JevBench v1.7.9: one same-draw self-hosted addition and one price correction
One new self-hosted row GLiClass v4 9B; Myles-G 12B now priced at its public list price; Super Decider links point to the vendor page; no method change. GLiClass answered 1,485 of 1,500 requests; 15 HTTP 422 length refusals count as misses. Its cost is an ESTIMATE from Qwen/Qwen3.5-9B with zero generated output. Myles-G 12B uses GBP 0.0015 per million input tokens, output free; prices exclude VAT. Price corrected in v1.7.9; earlier releases showed a same-base ESTIMATE of USD 0.0264 per 1,000. All other scores and confidence intervals are unchanged.
Aggregate results JSON · SHA-256 ab863afda444baa950b6007768d0048060e5661b019ac94b0a777d6dded63294 · Previous v1.7.8
JevBench v1.7.9 · headline
JevBench Capability Score
Capability ranking of Jev-class systems
Capability Score averages Intelligence and Calibration. kettei-fat leads the Jev-class systems with 86.3.
Jev-class means at most 2× the cost and median latency of Jev (v1.5 reference). How we choose ↘
Adjust cost / latency caps · 2× official
ScoreCCost
ScoreCap.Capability
Score$/1k$/1k decisions
- 1kettei-fatAPIv1.6.43 drawv1.7 · 10 s rule79.857.486.3$0.026*
- 2Super DeciderAPIv1.7.0 drawv1.7 · 10 s rule80.056.485.9$0.028
- 3Super DeciderAPIv1.7.0 drawv1.7 · 10 s rule78.964.384.9$0.016
- 4Fastino GLiDEAPIv1.6.43 drawv1.7 · 10 s rule76.159.184.5$0.023
- 5H2O-Lightning-31B v1.0.7Usev1.6.43 drawv1.7 · 10 s rule79.350.684.3$0.044*Use this model
- 6Drex MaxAPIv1.6.43 drawv1.7 · 10 s rule78.955.384.2$0.031
- 7basal-omni-1.0-ultraUsev1.6.43 drawv1.7 · 10 s rule74.150.482.1$0.045*Use this model
- 8DOPEST 1.0 31BUsev1.7.4 drawv1.7 · 10 s rule73.949.282.0$0.049*Use this model
- 9SubQ Snap 1.1APIv1.6.43 drawv1.7 · 10 s rule75.860.781.9$0.020*
- 10Jeff-1.1-MediumUsev1.7.8 drawv1.7 · 10 s rule74.553.881.9$0.035*Use this model
Show all 202 Jev-class systems (192 more)
- 11Arcbright Elf 1 31BUsev1.7.0 drawv1.7 · 10 s rule71.949.781.8$0.047*Use this model
- 12Quyet-1.0-LargeUsev1.7 · 10 s rule73.450.581.7$0.045*Use this model
- 13Jeff-1.0-LargeUsev1.6.4 drawv1.7 · 10 s rule71.949.281.5$0.049*Use this model
- 14SubQ SnapAPIv1.6.43 drawv1.7 · 10 s rule72.357.781.5$0.026*
- 15Sieve-31B-v2Usev1.7.5 drawv1.7 · 10 s rule68.750.981.1$0.043*Use this model
- 16d1-ultraAPIv1.6.43 drawv1.7 · 10 s rule69.762.380.5$0.018*
- 17Bespoke Nimble-v4-maxAPIv1.6.43 drawv1.7 · 10 s rule68.260.779.7$0.020*
- 18metask-jev-rain-31B-trained service v3Usev1.6.43 drawv1.7 · 10 s rule72.649.379.6$0.049*Use this model
- 19decisio v0.8.0 on gemma-4-31B-itUsev1.7 · 10 s rule70.451.779.6$0.041*Use this model
- 20metask-jev-rain-31B-trainedUsev1.6.43 drawv1.7 · 10 s rule70.750.778.9$0.044*Use this model
- 21Myles-G 31BAPIv1.6.43 drawv1.7 · 10 s rule73.651.178.8$0.043*
- 22Sage 1.3.0APIv1.7 · 10 s rule65.658.278.6$0.025*
- 23Bobcat 1.3Usev1.7.0 drawv1.7 · 10 s rule68.149.977.9$0.047*Use this model
- 24deck-31BUsev1.7 · 10 s rule73.049.777.6$0.048*Use this model
- 25Torchcast Decision 4BUsev1.6.43 drawv1.7 · 10 s rule67.164.877.5$0.015*Use this model
- 26Jev 1.13.0APIv1.7 · 10 s rule63.654.777.1$0.032
- 27metask-jev-rain-12B-trainedUsev1.6.43 drawv1.7 · 10 s rule66.552.776.9$0.038*Use this model
- 28Wald 26B-A4BUsev1.7.0 drawv1.7 · 10 s rule65.549.676.8$0.048*Use this model
- 29René-1 31B FP8Usev1.7 · 10 s rule61.746.076.0$0.063*Use this model
- 30Decisio 12B tuned commitUsev1.7.6 drawv1.7 · 10 s rule62.858.575.7$0.024*Use this model
- 31Jev 1.14.0APIv1.6.43 drawv1.7 · 10 s rule59.954.875.0$0.032*
- 32H2O-Lightning-4B v1.1Usev1.6.43 drawv1.7 · 10 s rule60.064.075.0$0.016*Use this model
- 33d1APIv1.7 · 10 s rule62.162.874.4$0.017
- 34SPX-CD FlashAPIv1.7 · 10 s rule58.452.174.3$0.039*
- 35Mercury DecideAPIv1.7 · 10 s rule65.962.173.6$0.018*
- 36Celeris-1 DecisionAPIv1.6.43 drawv1.7 · 10 s rule58.960.473.6$0.021*
- 37OpenAI DecisionsAPIv1.7 · 10 s rule56.948.673.5$0.052
- 38Surogate Rune 26B-A4B v3Usev1.7 · 10 s rule56.149.673.5$0.048*Use this model
- 39kettei-chubAPIv1.6.43 drawv1.7 · 10 s rule62.262.773.1$0.018*
- 40Bobcat Flash 1.2Usev1.7 · 10 s rule65.649.773.0$0.048*Use this model
- 41Torchcast Decision 4BUsev1.6.43 drawv1.7 · 10 s rule60.364.872.6$0.015*Use this model
- –Microsoft-Decision-1APIv1.6.43 drawv1.7 · 10 s rule59.262.272.0$0.018PRELIMINARY — API supplement A2 (300 sealed + 300 public), not the S1200 full set; equating pool 4 of 6
- 42torchcast-decision-12bUsev1.7 · 10 s rule60.556.471.7$0.028*Use this model
- 43decider-12b v2Usev1.7 · 10 s rule61.957.871.3$0.026*Use this model
- 44Jev-OmniUsev1.7 · 10 s rule55.556.171.3$0.029*Use this model
- 45Winnow-12B Q8Usev1.7 · 10 s rule59.556.671.2$0.028*Use this model
- 46Xor 26B-A4BUsev1.7 · 10 s rule57.550.771.2$0.044*Use this model
- 47decider-12b v1, stock Gemma-4-12B-itUsev1.7 · 10 s rule59.657.871.0$0.026*Use this model
- 48CygnetUsev1.7 · 10 s rule54.856.470.9$0.028*Use this model
- 49Microsoft-Decision-1APIv1.7 · 10 s rule57.261.570.8$0.019
- 50decisio v0.8.0 on gemma-4-12B-itUsev1.7 · 10 s rule52.359.370.7$0.023*Use this model
- 51Myles-G 12BAPIv1.6.43 drawv1.7 · 10 s rule57.399.470.7$0.0011*
- 52InstinctAPIv1.7 · 10 s rule47.864.470.5$0.015
- 53decider chat on Gemma-4-31B-itUsev1.7 · 10 s rule58.550.870.4$0.044*Use this model
- 54baselabs elephant-1Usev1.7.0 drawv1.7 · 10 s rule55.848.470.2$0.052*Use this model
- 55GEV-26B-DecideUsev1.7 · 10 s rule49.851.170.1$0.043*Use this model
- 56Chon-1.0-12B-FastUsev1.7.1 drawv1.7 · 10 s rule60.358.770.1$0.024*Use this model
- 57decisio v0.9.0 on gemma-4-12B-itUsev1.7 · 10 s rule52.359.370.0$0.023*Use this model
- 58Blink v0.3 26B-A4BUsev1.6.9 drawv1.7 · 10 s rule49.249.669.9$0.048*Use this model
- 59attest-31B-v1.0APIv1.6.43 drawv1.7 · 10 s rule65.250.769.5$0.044*
- 60Wald 4B v2.1Usev1.6.9 drawv1.7 · 10 s rule52.046.369.3$0.062*Use this model
- 61vLLM-SR d3-miniUsev1.7.0 drawv1.7 · 10 s rule50.864.268.8$0.016*Use this model
- 62Torchcast Decision 4BUsev1.6.43 drawv1.7 · 10 s rule54.264.868.7$0.015*Use this model
- 63Xevn-4BUsev1.7.0 drawv1.7 · 10 s rule52.166.068.6$0.014*Use this model
- 64Chon-1.0-E4BUsev1.7.3 drawv1.7 · 10 s rule55.570.768.5$0.0094*Use this model
- 65Hopper 12B trainedUsev1.7 · 10 s rule51.756.268.1$0.029*Use this model
- 66SPX-CD-OmniUsev1.7 · 10 s rule48.858.368.1$0.025*Use this model
- 67Aplomb 1Usev1.7 · 10 s rule45.257.967.1$0.025*Use this model
- 68Bespoke Nimble 9B v3Usev1.7 · 10 s rule56.546.166.8$0.063*Use this model
- 69Amazon Strands Decider 26B-A4BUsev1.7.1 drawv1.7 · 10 s rule46.750.566.7$0.045*Use this model
- 70Metask rain 4BUsev1.6.4 drawv1.7 · 10 s rule51.151.866.0$0.040*Use this model
- 71shisa-de-1Usev1.6.43 drawv1.7 · 10 s rule49.949.865.5$0.047*Use this model
- 72Plumb-4BUsev1.7 · 10 s rule42.863.165.3$0.017*Use this model
- 73Arcbright Elf 1 12BUsev1.6.43 drawv1.7 · 10 s rule43.857.465.0$0.026*Use this model
- 74decider-4b v2Usev1.7 · 10 s rule40.164.564.9$0.015*Use this model
- 75Seb-9B-v2Usev1.7.0 drawv1.7 · 10 s rule46.148.964.8$0.051*Use this model
- 76HS1 0.1.2APIv1.6.43 drawv1.7 · 10 s rule52.949.164.3$0.050*
- 77JevK5 v0.3Usev1.7 · 10 s rule38.563.164.2$0.017*Use this model
- 78Bespoke Nimble-v4APIv1.6.43 drawv1.7 · 10 s rule41.460.764.0$0.020*
- 79Jeffan-4B-2010Usev1.7.0 drawv1.7 · 10 s rule40.164.164.0$0.016*Use this model
- 80Clef-FlashUsev1.7 · 10 s rule42.246.863.8$0.059*Use this model
- 81JPT-4BUsev1.6.43 drawv1.7 · 10 s rule40.864.263.8$0.016*Use this model
- 82Vansa-3.4APIv1.7 · 10 s rule40.961.963.7$0.019
- 83deck-4B v1.0Usev1.7 · 10 s rule38.663.563.6$0.017*Use this model
- 84Jebadiah 9B v2.1Usev1.6.43 drawv1.7 · 10 s rule37.248.963.4$0.051*Use this model
- 85GLiClass v4 9BUsev1.7.9 drawv1.7 · 10 s rule39.651.263.2$0.042*Use this model
- 86Seb-9BUsev1.6.43 drawv1.7 · 10 s rule38.348.963.1$0.051*Use this model
- 87janus 4BUsev1.7 · 10 s rule37.664.462.9$0.015*Use this model
- 88Amos 1 mediumAPIv1.6.43 drawv1.7 · 10 s rule38.148.762.9$0.051*
- 89JevK5 v0.2.0Usev1.7 · 10 s rule36.363.162.8$0.017*Use this model
- 90Messier One v0.4Usev1.6.43 drawv1.7 · 10 s rule52.061.162.4$0.020*Use this model
- 91Imajev-4BUsev1.7 · 10 s rule34.663.361.9$0.017*Use this model
- 92jqvUsev1.7 · 10 s rule33.951.261.9$0.042*Use this model
- 93metask-jev-4bUsev1.7 · 10 s rule39.257.761.8$0.026*Use this model
- 94Kahn1 4BUsev1.6.43 drawv1.7 · 10 s rule39.750.161.7$0.046*Use this model
- 95pplx-decider-flash-v1.2Usev1.6.43 drawv1.7 · 10 s rule45.564.461.7$0.015*Use this model
- 96levUsev1.7 · 10 s rule41.161.861.7$0.019*Use this model
- 97Decision 2.0 Lux 9BUsev1.6.43 drawv1.7 · 10 s rule35.848.561.6$0.052*Use this model
- 98blink-4bUsev1.6.43 drawv1.7 · 10 s rule35.063.561.4$0.016*Use this model
- 99HopperUsev1.7 · 10 s rule33.662.361.2$0.018*Use this model
- 100Malkuth-4BUsev1.7 · 10 s rule33.855.861.1$0.030*Use this model
- 101Decision 4B v1.2Usev1.7 · 10 s rule32.463.160.4$0.017*Use this model
- 102jiwo 4bUsev1.6.9 drawv1.7 · 10 s rule34.864.160.2$0.016*Use this model
- 103Verdin-4BUsev1.7.0 drawv1.7 · 10 s rule35.462.259.6$0.018*Use this model
- 104Wald 4BUsev1.6.4 drawv1.7 · 10 s rule41.054.359.6$0.033*Use this model
- 105Quyet-1.0-MediumUsev1.7 · 10 s rule41.763.459.2$0.017*Use this model
- 106Decision-4BUsev1.7 · 10 s rule37.965.959.1$0.014*Use this model
- 107Manchego v2.1Usev1.7 · 10 s rule32.564.358.3$0.015*Use this model
- 108Messier One v0.2Usev1.7 · 10 s rule42.464.858.1$0.015Use this model
- 109Diffusion JevUsev1.7 · 10 s rule54.949.658.1$0.048*Use this model
- 110jev-localUsev1.7 · 10 s rule41.958.957.5$0.024*Use this model
- 111Instinct Dual 4BAPIv1.7 · 10 s rule28.375.057.2$0.0068
- 112Decision 4B v1.1Usev1.7 · 10 s rule29.163.157.2$0.017*Use this model
- 113JEV Qwen3.5-9B Base NVFP4Usev1.7 · 10 s rule29.347.657.1$0.056*Use this model
- 114iota-decide-4bUsev1.6.43 drawv1.7 · 10 s rule29.463.057.1$0.017*Use this model
- 115Uniphore-4B-DeciderUsev1.6.43 drawv1.7 · 10 s rule34.264.557.1$0.015*Use this model
- 116Jebadiah 9B v2Usev1.6.43 drawv1.7 · 10 s rule30.648.956.9$0.051*Use this model
- 117CoCo-Decision-4BUsev1.7 · 10 s rule32.765.256.9$0.014*Use this model
- 118APUS-OpenJev-v1-9BUsev1.7 · 10 s rule47.248.756.9$0.051*Use this model
- 119Decider-4BUsev1.6.43 drawv1.7 · 10 s rule27.166.256.8$0.013*Use this model
- 120typecastlmUsev1.7 · 10 s rule29.864.056.5$0.016*Use this model
- 121pjev-2bUsev1.7.0 drawv1.7 · 10 s rule28.863.556.5$0.016*Use this model
- 122TypeCastLM 1.4.0Usev1.7 · 10 s rule35.464.856.4$0.015*Use this model
- 123RYOTIDE-QwenUsev1.6.43 drawv1.7 · 10 s rule25.262.256.3$0.018*Use this model
- 124Compass 0.2.0Usev1.6.43 drawv1.7 · 10 s rule28.957.656.1$0.026*Use this model
- 125Standard One 3B SH v4Usev1.7.0 drawv1.7 · 10 s rule34.448.656.0$0.052*Use this model
- 126Standard One 3B SHUsev1.6.43 drawv1.7 · 10 s rule32.848.655.6$0.052*Use this model
- 127SemIfUsev1.7 · 10 s rule27.163.155.4$0.017*Use this model
- 128ArseneLupin v1.1Usev1.6.43 drawv1.7 · 10 s rule28.060.755.1$0.020*Use this model
- 129local-jev Qwen3.5-4BUsev1.7 · 10 s rule24.159.355.0$0.023*Use this model
- 130Decision 2BUsev1.7 · 10 s rule22.966.454.8$0.013*Use this model
- 131Jez-large-1.0.0Usev1.7.0 drawv1.7 · 10 s rule25.670.554.4$0.0096*Use this model
- 132NomaUsev1.6.43 drawv1.7 · 10 s rule25.162.054.3$0.019*Use this model
- 133spark-s1-4b-v6Usev1.7 · 10 s rule43.260.454.0$0.021*Use this model
- 134Imajev-2BUsev1.6.43 drawv1.7 · 10 s rule21.251.854.0$0.041*Use this model
- 135RYOTIDE-Qwen9Usev1.6.4 drawv1.7 · 10 s rule27.649.452.5$0.049*Use this model
- 136classify-1APIv1.7.3 drawv1.7 · 10 s rule21.850.852.3$0.044
- 137Arbiter-4BUsev1.6.43 drawv1.7 · 10 s rule27.466.252.3$0.013*Use this model
- 138DecisionNode Edge 0.9.2Usev1.7.0 drawv1.7 · 10 s rule26.871.252.1$0.0091*Use this model
- 139system-one-openAPIv1.7 · 10 s rule27.968.351.9$0.011*
- 140D1A on Gemma 4 E4BUsev1.7.7 drawv1.7 · 10 s rule19.960.651.1$0.021*Use this model
- 141ClassOne Qwen 3.5 9BUsev1.7 · 10 s rule42.249.350.5$0.049*Use this model
- 142Liquid AI d1-3BUsev1.6.9 drawv1.7 · 10 s rule22.150.649.5$0.044*Use this model
- 143ZeroEntropy zerank-2Usev1.7 · 10 s rule15.648.449.4$0.052*Use this model
- 144open-alternative-jevUsev1.7 · 10 s rule22.363.349.1$0.017*Use this model
- 145OpenSourceJevUsev1.7 · 10 s rule23.269.148.6$0.011*Use this model
- 146EXAONE-4.0-1.2B-JEV v0.3Usev1.7 · 10 s rule14.857.947.7$0.025*Use this model
- 147Malkuth-2BUsev1.7 · 10 s rule15.965.647.6$0.014*Use this model
- 148DecisionNode Edge Mini 0.9.2Usev1.7.0 drawv1.7 · 10 s rule17.680.247.6$0.0046*Use this model
- 149Nemotron Diffusion 8BUsev1.7 · 10 s rule20.455.547.5$0.030*Use this model
- 150jiwo 0.8bUsev1.6.9 drawv1.7 · 10 s rule15.978.447.2$0.0052*Use this model
- 151decision-machine-1APIv1.7 · 10 s rule13.158.346.1$0.024
- 152decisor-4bUsev1.6.4 drawv1.7 · 10 s rule39.664.045.7$0.016*Use this model
- 153D1A on Gemma 4 E2BUsev1.7.7 drawv1.7 · 10 s rule8.960.645.5$0.021*Use this model
- 154Qwen3-Reranker-4BUsev1.7 · 10 s rule16.348.444.7$0.052*Use this model
- 155decider-2bUsev1.7 · 10 s rule20.764.944.6$0.015*Use this model
- 156BAAI bge-reranker-v2-m3Usev1.7 · 10 s rule0.159.344.1$0.023*Use this model
- 157kev 4BUsev1.7 · 10 s rule19.565.844.0$0.014*Use this model
- 158Mixedbread mxbai-rerank-base-v2Usev1.7 · 10 s rule1.160.343.9$0.021*Use this model
- 159lev-350mUsev1.7 · 10 s rule4.880.043.3$0.0046*Use this model
- 160Jez-small-1.0.0Usev1.7.0 drawv1.7 · 10 s rule4.679.743.2$0.0048*Use this model
- 161Certo v1Usev1.7 · 10 s rule0.296.943.2$0.0013*Use this model
- 162PurpleMIST-Mini-1.0Usev1.7.0 drawv1.7 · 10 s rule8.470.843.0$0.0094*Use this model
- 163Chon-1.0-12BUsev1.6.43 drawv1.7 · 10 s rule27.349.043.0$0.050*Use this model
- 164smalljev semantic-v9Usev1.7 · 10 s rule8.760.843.0$0.020*Use this model
- 165Vega-UUsev1.7.0 drawv1.7 · 10 s rule8.564.942.6$0.015*Use this model
- 166jul fastUsev1.7 · 10 s rule7.569.042.6$0.011*Use this model
- 167Krino-2BUsev1.6.43 drawv1.7 · 10 s rule22.061.042.4$0.020*Use this model
- 168Alibaba GTE Reranker ModernBERT-baseUsev1.7 · 10 s rule0.069.341.7$0.011*Use this model
- 169Jeetu Decider 2BUsev1.7.2 drawv1.7 · 10 s rule10.270.341.7$0.0098*Use this model
- 170watt-flash-0.1Usev1.7 · 10 s rule3.889.041.6$0.0023Use this model
- 171Decision FastUsev1.7 · 10 s rule6.180.140.6$0.0046*Use this model
- 172openJev Verdict 1.4Usev1.7 · 10 s rule5.086.640.4$0.0028*Use this model
- 173Quyet-1.0-Small-ENUsev1.7 · 10 s rule6.189.139.5$0.0023*Use this model
- 174ClassOne Gemma 4 E4BUsev1.6.9 drawv1.7 · 10 s rule28.269.338.4$0.011*Use this model
- 175ClassOne Qwen 3.5 4BUsev1.6.9 drawv1.7 · 10 s rule27.365.038.1$0.015*Use this model
- 176OpenDecisionUsev1.7 · 10 s rule7.479.138.0$0.0050*Use this model
- 177verdict-smallUsev1.7 · 10 s rule2.3100.037.6$0.00087*Use this model
- 178kev 0.6BUsev1.7 · 10 s rule7.680.136.9$0.0046*Use this model
- 179Vega 4BUsev1.6.9 drawv1.7 · 10 s rule15.747.236.6$0.058*Use this model
- 180WaterSheepUsev1.7 · 10 s rule5.368.735.5$0.011*Use this model
- 181Raw Phi-4 mini direct logitsUsev1.7 · 10 s rule13.753.235.3$0.036*Use this model
- 182Fastino GLiNER-2.5-DecideAPIv1.7 · 10 s rule7.045.834.6$0.064
- 183Raw Qwen3 4B Instruct 2507 direct logitsUsev1.7 · 10 s rule34.063.434.6$0.017*Use this model
- 184Vega 0.8BUsev1.6.9 drawv1.7 · 10 s rule3.161.533.0$0.019*Use this model
- 185kev 0.5BUsev1.7 · 10 s rule4.680.132.9$0.0046*Use this model
- 186Liquid AI d1-omni-600MUsev1.6.9 drawv1.7 · 10 s rule3.373.132.8$0.0079*Use this model
- 187Quyet-1.0-SmallUsev1.7 · 10 s rule4.079.632.5$0.0048*Use this model
- 188Quyet-1.0-TinyUsev1.7 · 10 s rule3.788.431.1$0.0024*Use this model
- 189Open Jev JSON CanvasUsev1.7 · 10 s rule61.249.330.6$0.049*Use this model
- 190Phocinae-Largha-150M-v1Usev1.6.43 drawv1.7 · 10 s rule3.695.330.2$0.0014*Use this model
- 191JevActAPIv1.7 · 10 s rule10.168.628.8$0.011*
- 192openJev VerdictUsev1.7 · 10 s rule4.186.626.6$0.0028*Use this model
- 193Tacet SonataUsev1.7 · 10 s rule2.289.024.2$0.0023*Use this model
- 194CLM-8BUsev1.7 · 10 s rule0.150.520.1$0.045*Use this model
- 195ClassOne Qwen 3.5 2BUsev1.6.9 drawv1.7 · 10 s rule5.870.219.2$0.0098*Use this model
- 196Deem 0.8B v1Usev1.7 · 10 s rule6.580.318.1$0.0045*Use this model
- 197Laya multilingualUsev1.7 · 10 s rule0.382.217.0$0.0039*Use this model
- 198ClassOne Gemma 4 E2BUsev1.7 · 10 s rule7.870.215.2$0.0098*Use this model
- 199Julia-1Usev1.6.43 drawv1.7 · 10 s rule2.489.79.9$0.0022*Use this model
- 200Raw Qwen3 1.7B direct logitsUsev1.7 · 10 s rule5.168.58.3$0.011*Use this model
- 201Raw Qwen3 0.6B direct logitsUsev1.7 · 10 s rule5.677.68.2$0.0056*Use this model
Wide coloured bar = Capability Score (0–100). Thin bars: cost above, median latency below; shared log ratio scale 0.25× … 64× Jev (v1.5 reference), ticks at 1× and 2× (cap). Shorter is cheaper or faster. * = est. (estimated cost). # counts ranked Jev-class systems; “–” marks unranked or outside systems. Tap ⓘ or hover a row or thin bar for details.
Cost and latency correlate here (Spearman ρ = 0.28, n = 282).
- Jev — reference (TypeSafe, closed) (2)
- Closed API (weights not public) (32)
- Open weights · LLM decoder (208)
- Open weights · diffusion LM (6)
- Open weights · encoder / classifier (21)
- Open weights · reranker (5)
- Base model control (no decision fine-tune, raw logits) (5)
- System (router / cascade / ensemble) (3)
- green: ≤ reference
- amber: ≤ cap (2× reference)
- red: > cap
Use = run this open-weights model yourself: own endpoint, API, local install or hardware, on Decision Models. Decision Models is run by the same company as Benchmark Heaven. Every open-weights row gets the same link, whether or not we serve the model; it never affects scores or ranks. Independence
Show general-purpose LLMs and other systems outside the limits
Sorted by Capability Score, not numbered. Each row says which limit it misses, measured against Jev ($0.032 per 1,000 decisions, median 0.62 s).
- –GPT-6 Luna (medium)APIv1.7 · 10 s rule96.938.498.4$0.11Outside: cost 3.5× Jev (v1.5 reference), latency 3.1× Jev (v1.5 reference)
- –GPT-6 Luna (low)APIv1.7 · 10 s rule95.739.097.1$0.11Outside: cost 3.3× Jev (v1.5 reference), latency 3.1× Jev (v1.5 reference)
- –GPT-5.6 LunaAPIv1.7 · 10 s rule93.030.293.6$0.21Outside: cost 6.6× Jev (v1.5 reference), latency 2.6× Jev (v1.5 reference)
- –DeepSeek V4.1 FlashAPIv1.7 · 10 s rule84.519.689.6$0.48Outside: cost 14.9× Jev (v1.5 reference), latency 3.6× Jev (v1.5 reference)
- –baselabs mammoth-1Usev1.7.0 drawv1.7 · 10 s rule76.239.585.2$0.10*Use this modelOutside: cost 3.2× Jev (v1.5 reference)
- –KnowLine 31B-Gen1Usev1.7.0 drawv1.7 · 10 s rule77.440.985.2$0.093*Use this modelOutside: cost 2.9× Jev (v1.5 reference)
- –Fastino GLiDE 1.2Usev1.7.0 drawv1.7 · 10 s rule76.430.684.9$0.21*Use this modelOutside: cost 6.4× Jev (v1.5 reference)
- –Perplexity Decider v1.1 27BUsev1.7 · 10 s rule75.629.982.7$0.22*Use this modelOutside: cost 6.7× Jev (v1.5 reference)
- –vLLM-SR d3-turboUsev1.6.43 drawv1.7 · 10 s rule74.829.882.6$0.22*Use this modelOutside: cost 6.8× Jev (v1.5 reference)
- –StartLux-Decision v1.1 27B InstantUsev1.6.43 drawv1.7 · 10 s rule73.230.482.6$0.21*Use this modelOutside: cost 6.5× Jev (v1.5 reference)
- –vLLM-SR d3Usev1.6.43 drawv1.7 · 10 s rule72.929.882.1$0.22*Use this modelOutside: cost 6.8× Jev (v1.5 reference)
- –Jeffan-31b-1710Usev1.7.0 drawv1.7 · 10 s rule71.642.881.7$0.081*Use this modelOutside: cost 2.5× Jev (v1.5 reference)
- –AJev lm6qUsev1.7.0 drawv1.7 · 10 s rule71.541.081.4$0.093*Use this modelOutside: cost 2.9× Jev (v1.5 reference)
- –Nimble-v4-maxUsev1.7.0 drawv1.7 · 10 s rule71.030.181.2$0.21*Use this modelOutside: cost 6.6× Jev (v1.5 reference)
- –Turion Jev 31BUsev1.7.1 drawv1.7 · 10 s rule70.940.680.9$0.096*Use this modelOutside: cost 3.0× Jev (v1.5 reference)
- –pplx-decider-v1.2Usev1.6.43 drawv1.7 · 10 s rule71.730.080.8$0.21*Use this modelOutside: cost 6.6× Jev (v1.5 reference)
- –torchcast-decision-27bUsev1.7 · 10 s rule70.430.380.2$0.21*Use this modelOutside: cost 6.5× Jev (v1.5 reference)
- –Vera LargeUsev1.7.0 drawv1.7 · 10 s rule67.829.979.6$0.22*Use this modelOutside: cost 6.7× Jev (v1.5 reference)
- –Kev 27BUsev1.7 · 10 s rule64.218.378.9$0.53*Use this modelOutside: cost 16.4× Jev (v1.5 reference)
- –StartLux-Decision v1.1 27B AutoThinkUsev1.6.43 drawv1.7 · 10 s rule66.723.578.8$0.35*Use this modelOutside: cost 11.0× Jev (v1.5 reference)
- –SPX-CD ProAPIv1.7 · 10 s rule64.643.178.5$0.079*Outside: cost 2.4× Jev (v1.5 reference)
- –Amos 1 largeAPIv1.6.43 drawv1.7 · 10 s rule68.230.078.0$0.21*Outside: cost 6.6× Jev (v1.5 reference)
- –Eikos-27BUsev1.7 · 10 s rule64.428.777.2$0.24*Use this modelOutside: cost 7.4× Jev (v1.5 reference)
- –Fastino GLiDEAPIv1.7 · 10 s rule66.135.977.0$0.14Outside: cost 4.3× Jev (v1.5 reference)
- –wity-1APIv1.7 · 10 s rule67.958.876.9$0.024Outside: latency 3.0× Jev (v1.5 reference)API price · eligibility checked at the developer's list price; at base-model pricing it would exceed the cost cap (2.70× Jev)
- –Aleria Judge 1.1.1Usev1.7.0 drawv1.7 · 10 s rule68.035.876.9$0.14*Use this modelOutside: cost 4.3× Jev (v1.5 reference)
- –KnowLine 35B-A3B-Gen1.5Usev1.7.0 drawv1.7 · 10 s rule60.324.776.7$0.32*Use this modelOutside: cost 10.0× Jev (v1.5 reference)
- –OpenJevUsev1.7 · 10 s rule70.228.575.8$0.24*Use this modelOutside: cost 7.5× Jev (v1.5 reference), latency 3.1× Jev (v1.5 reference)
- –ClefUsev1.7 · 10 s rule59.128.175.3$0.25*Use this modelOutside: cost 7.7× Jev (v1.5 reference)
- –SimpleJev Qwen3.8-27BUsev1.7 · 10 s rule57.915.174.9$0.68*Use this modelOutside: cost 20.9× Jev (v1.5 reference), latency 2.3× Jev (v1.5 reference)
- –JEV-27BUsev1.7 · 10 s rule56.831.073.9$0.20*Use this modelOutside: cost 6.2× Jev (v1.5 reference)
- –Nagi-ENORMOUS 27BUsev1.6.43 drawv1.7 · 10 s rule62.430.073.8$0.22*Use this modelOutside: cost 6.7× Jev (v1.5 reference)
- –RSI-Jev v6.1-VL 27BUsev1.6.9 drawv1.7 · 10 s rule61.631.173.5$0.20*Use this modelOutside: cost 6.1× Jev (v1.5 reference)
- –Sage 1.4APIv1.6.43 drawv1.7 · 10 s rule63.759.573.3$0.022*Outside: latency 2.8× Jev (v1.5 reference)
- –Autoloops – Gemma 4 31B ITAPIv1.7 · 10 s rule58.140.572.7$0.096Outside: cost 3.0× Jev (v1.5 reference)
- –swanOneUsev1.7 · 10 s rule56.242.272.4$0.085*Use this modelOutside: cost 2.6× Jev (v1.5 reference)
- –FrontiersMind Lumma-fev-27BUsev1.7.0 drawv1.7 · 10 s rule55.517.972.3$0.54*Use this modelOutside: cost 16.9× Jev (v1.5 reference)
- –AutoJev-27BUsev1.7 · 10 s rule55.629.471.6$0.23*Use this modelOutside: cost 7.0× Jev (v1.5 reference)
- –SimpleJev Qwen3.8-27BAPIv1.7 · 10 s rule52.914.971.6$0.69*Outside: cost 21.3× Jev (v1.5 reference), latency 2.1× Jev (v1.5 reference)
- –Decision 2.0 Vega 27BUsev1.7 · 10 s rule60.429.671.3$0.22*Use this modelOutside: cost 6.9× Jev (v1.5 reference)
- –JPT-35B-A3BUsev1.7 · 10 s rule51.433.670.7$0.16*Use this modelOutside: cost 5.0× Jev (v1.5 reference)
- –Jebadiah 27BUsev1.7 · 10 s rule57.529.970.2$0.22*Use this modelOutside: cost 6.7× Jev (v1.5 reference)
- –NInfer Qwen3.8-Flash-Next mixedUsev1.7 · 10 s rule49.042.569.7$0.082*Use this modelOutside: cost 2.5× Jev (v1.5 reference)
- –Respan Span-02Usev1.6.43 drawv1.7 · 10 s rule60.945.169.3$0.068*Use this modelOutside: cost 2.1× Jev (v1.5 reference)
- –JevOneUsev1.7 · 10 s rule46.039.868.9$0.10*Use this modelOutside: cost 3.1× Jev (v1.5 reference)
- –SimpleJev Qwen3.6-35B-A3BAPIv1.7 · 10 s rule56.335.168.3$0.15*Outside: cost 4.5× Jev (v1.5 reference), latency 2.1× Jev (v1.5 reference)
- –NInfer Qwen3.8-27B NVFP4Usev1.7 · 10 s rule51.829.168.0$0.23*Use this modelOutside: cost 7.1× Jev (v1.5 reference)
- –Open-Jev 27B v1.1Usev1.7 · 10 s rule52.514.468.0$0.72*Use this modelOutside: cost 22.2× Jev (v1.5 reference), latency 2.9× Jev (v1.5 reference)
- –LitJevUsev1.7 · 10 s rule43.328.466.6$0.24*Use this modelOutside: cost 7.6× Jev (v1.5 reference), latency 5.1× Jev (v1.5 reference)
- –Tern-35B-A3BUsev1.7.0 drawv1.7 · 10 s rule44.142.166.4$0.085*Use this modelOutside: cost 2.6× Jev (v1.5 reference)
- –decider-35b-a3bUsev1.7 · 10 s rule48.734.466.2$0.15*Use this modelOutside: cost 4.8× Jev (v1.5 reference)
- –Bev / Bonsai 27BUsev1.7 · 10 s rule43.828.266.1$0.25*Use this modelOutside: cost 7.6× Jev (v1.5 reference), latency 3.5× Jev (v1.5 reference)
- –Imajev-9BUsev1.6.43 drawv1.7 · 10 s rule36.730.862.4$0.20*Use this modelOutside: cost 6.3× Jev (v1.5 reference)
- –Open-Jev 9BUsev1.7 · 10 s rule43.933.161.8$0.17*Use this modelOutside: cost 5.3× Jev (v1.5 reference), latency 2.3× Jev (v1.5 reference)
- –JEB-35B-A3BUsev1.6.43 drawv1.7 · 10 s rule33.242.961.6$0.080*Use this modelOutside: cost 2.5× Jev (v1.5 reference)
- –Qwen3.5-9B Jev-like data-mix v2Usev1.7 · 10 s rule41.945.760.9$0.065*Use this modelOutside: cost 2.001× Jev (v1.5 reference)
- –Standard One 8B SHUsev1.6.9 drawv1.7 · 10 s rule34.543.260.0$0.078*Use this modelOutside: cost 2.4× Jev (v1.5 reference)
- –Gemini 3.1 Flash-LiteAPIv1.7 · 10 s rule58.629.259.8$0.23Outside: cost 7.1× Jev (v1.5 reference)
- –reflex 4BUsev1.7 · 10 s rule30.863.158.9$0.017*Use this modelOutside: latency 4.7× Jev (v1.5 reference)
- –openjev-sglangAPIv1.7 · 10 s rule38.835.658.7$0.14*Outside: cost 4.3× Jev (v1.5 reference), latency 2.1× Jev (v1.5 reference)
- –Standard One 8BUsev1.7 · 10 s rule32.843.358.4$0.078*Use this modelOutside: cost 2.4× Jev (v1.5 reference)
- –Clef-omniUsev1.6.9 drawv1.7 · 10 s rule34.526.858.3$0.28*Use this modelOutside: cost 8.5× Jev (v1.5 reference)
- –Bespoke Nimble 9BUsev1.7 · 10 s rule43.136.856.5$0.13*Use this modelOutside: cost 4.0× Jev (v1.5 reference)
- –kev 8BUsev1.7 · 10 s rule29.340.450.0$0.097*Use this modelOutside: cost 3.0× Jev (v1.5 reference)
- –Nev-2BUsev1.6.43 drawv1.6 · re-measure pending20.271.049.9$0.0092*Use this modelOutside: latency 20.8× Jev (v1.5 reference)
- –JADEUsev1.7 · 10 s rule47.933.649.2$0.16*Use this modelOutside: cost 5.1× Jev (v1.5 reference)
- –Qwen3.8 27BAPIv1.7 · 10 s rule33.40.048.6$2.18*Outside: cost 67.4× Jev (v1.5 reference), latency 13.5× Jev (v1.5 reference)
- –Open-Jev 2BUsev1.7 · 10 s rule21.833.147.7$0.17*Use this modelOutside: cost 5.3× Jev (v1.5 reference)
- –Laya typed-decisionsUsev1.7 · 10 s rule5.582.544.2$0.0038*Use this modelOutside: latency 3.5× Jev (v1.5 reference)
- –Qwen3.5-0.8B Decision ModelUsev1.6 · re-measure pending7.379.542.6$0.0048*Use this modelOutside: latency 18.3× Jev (v1.5 reference)
- –Gutsy 0.8B v0.3Usev1.6.9 drawv1.7 · 10 s rule8.678.442.2$0.0053*Use this modelOutside: latency 3.1× Jev (v1.5 reference)
- –Reflex Instinct 0.6BUsev1.6.43 drawv1.6 · re-measure pending4.079.838.6$0.0047*Use this modelOutside: latency 6.3× Jev (v1.5 reference)
- –Vega-UUsev1.6.43 drawv1.7 · 10 s rule5.379.236.1$0.0049*Use this modelOutside: latency 2.6× Jev (v1.5 reference)
- –open-jev-deberta-v3-largeUsev1.7 · 10 s rule2.877.635.4$0.0056*Use this modelOutside: latency 5.7× Jev (v1.5 reference)
- –modernbert-jev v2Usev1.6.9 drawv1.6 · re-measure pending2.670.434.6$0.0097*Use this modelOutside: latency 8.7× Jev (v1.5 reference)
- –LayaUsev1.7 · 10 s rule1.884.932.5$0.0032*Use this modelOutside: latency 3.3× Jev (v1.5 reference)
- –SimpleJevUsev1.7 · 10 s rule12.770.331.1$0.0098*Use this modelOutside: latency 14.9× Jev (v1.5 reference)
- –Raw Qwen3 8B direct logitsUsev1.7 · 10 s rule26.545.530.1$0.065*Use this modelOutside: cost 2.02× Jev (v1.5 reference)
- –system-oneUsev1.7 · 10 s rule29.345.029.7$0.068*Use this modelOutside: cost 2.1× Jev (v1.5 reference)
- –MirrorUsev1.7 · 10 s rule0.089.325.8$0.0023*Use this modelOutside: latency 2.1× Jev (v1.5 reference)
Jev-class = cost per decision at most 2× Jev's (≤ $0.065 per 1,000 decisions) and median latency at most 2× Jev's (≤ 1.23 s), the adjusted p50 — the same median the speed chart plots, not the four-axis Speed score. 202 of 282 systems qualify; the other 80, including the general-purpose LLMs, are listed below the divider in the ranking. The charts below show speed and cost beside Capability Score; the official JevBench Score weighs all four axes.
Capability against cost and speed
Jev-class systems are shown by default. Bubble size follows the official JevBench Score; official rank stays unchanged. The five most capable Jev-class systems are labelled.
Capability vs cost
Upper right is best: more capable and cheaper. Cost is USD per 1,000 decisions on a log scale. The dashed line is 2× Jev’s cost.
Capability vs speed
Upper right is best: more capable and faster. Speed here is the median-latency speed — the same adjusted median (p50) latency the Jev-class limit uses — a log scale, so each 20 points is 10× faster (median under the numbers). Self-hosted systems are plotted at the est. incl. network latency (2x measured + 150 ms); their measured loopback p50 (on-GPU, no network; not directly comparable with API latency) is in the tooltip. The dashed line is 2× Jev’s median latency.
- Jev — reference (TypeSafe, closed) (2)
- Closed API (weights not public) (31)
- Open weights · LLM decoder (208)
- Open weights · diffusion LM (6)
- Open weights · encoder / classifier (21)
- Open weights · reranker (5)
- Base model control (no decision fine-tune, raw logits) (5)
- System (router / cascade / ensemble) (3)
- faint = outside Jev-class
Model kind
Jev-class
Release
No system in this release reports an exact parameter count.
JevBench v1.7.9
JevBench Composite Score: 277 ranked systems
Official· four axes 0–100, equal-weight harmonic mean · Method notes ↓
Adjust weights ↓
The official order weighs Intelligence, Calibration, Speed and Cost equally. Each button re-sorts the same systems by one axis, and the column headings sort too.
Greener = stronger within its column.
282 of 282 systems, sorted by official rank, #1 first.
- 1Super DeciderAPIfine-tune · 26B-A4B · NVFP4Base model: google/diffusiongemma-26B-A4B-itmodel card*178.5I 79C 91S 85K 64$0.016
- 2Myles-G 12BAPIClosed APIBase model: google/gemma-4-12B-itmodel card*278.2I 57C 84S 85K 99est.$0.0011
- 3kettei-fatAPIClosed APIBase model: undisclosed77.7I 80C 93S 93K 57est.$0.026
- 4Fastino GLiDE (hosted API, 22 Oct 2026)APIClosed APIBase model: undisclosed*376.4I 76C 93S 87K 59$0.023
- 5Torchcast Decision 4B (v1.3)Usefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource76.2I 67C 88S 92K 65est.$0.015
- 6d1-ultraAPIClosed APIBase model: undisclosed76.0I 70C 91S 89K 62est.$0.018
- 7SubQ Snap 1.1APIClosed APIBase model: undisclosed75.7I 76C 88S 85K 61est.$0.020
- 8Drex MaxAPIClosed APIBase model: Qwen3.8-27Bvendor statement*475.4I 79C 90S 90K 55$0.031
- 9Bespoke Nimble-v4-maxAPIClosed APIBase model: Qwen/Qwen3.8-27Bmodel card*574.8I 68C 91S 88K 61est.$0.020
- 10Super Decider (thinking)APIfine-tune · 26B-A4B · NVFP4Base model: google/diffusiongemma-26B-A4B-itmodel card*674.5I 80C 92S 80K 56$0.028
- 11SubQ SnapAPIClosed APIBase model: undisclosed74.4I 72C 91S 86K 58est.$0.026
- 12H2O-Lightning-4B v1.1Usefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource73.8I 60C 90S 92K 64est.$0.016
- 13Sage 1.3.0APIClosed APIBase model: undisclosed73.7I 66C 92S 92K 58est.$0.025
- 14kettei-chubAPIClosed APIBase model: undisclosed73.4I 62C 84S 95K 63est.$0.018
- 15H2O-Lightning-31B v1.0.7Usefine-tune · 31BBase model: google/gemma-4-31B-itsource73.1I 79C 89S 90K 51est.$0.044
- 16Torchcast Decision 4B (v1.2)Usefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource73.0I 60C 85S 91K 65est.$0.015
- 17d1APIClosed APIBase model: undisclosed72.8I 62C 87S 88K 63$0.017
- 18Mercury DecideAPISystemBase model: undisclosedsource72.2I 66C 81S 85K 62est.$0.018
- 19Decisio 12B tuned commitUsemerge · 12B · BF16Base model: google/gemma-4-12B-itsource72.2I 63C 89S 90K 58est.$0.024
- 20Chon-1.0-E4BUsefine-tune · 4BBase model: google/gemma-4-E4B-itsource71.8I 55C 82S 89K 71est.$0.0094
Show all 282 systems (257 more ranked, 5 more not ranked)
- 21decisio v0.8.0 on gemma-4-31B-itUse31B · FP8 on loadBase model: google/gemma-4-31B-itsource71.7I 70C 89S 91K 52est.$0.041
- 22basal-omni-1.0-ultraUsemerge · 31BBase model: google/gemma-4-31B-itsource71.5I 74C 90S 87K 50est.$0.045
- 23Jev 1.13.0APIJev referenceBase model: undisclosed71.4I 64C 91S 91K 55$0.032
- Microsoft-Decision-1 (Azure Foundry) — A2 re-run (preliminary)APIPRELIMINARY — API supplement A2 (300 sealed + 300 public), not the S1200 full set; equating pool 4 of 6Closed APIBase model: Qwen3.5-9Bvendor statementvendor statement · Base model disclosed by Microsoft; the Decision-1 endpoint is evaluated as a hosted API offering.71.4I 59C 85S 89K 62$0.018
- 24Quyet-1.0-LargeUseadapter · 31BBase model: google/gemma-4-31B-itsource71.3I 73C 90S 86K 51est.$0.045
- 25Celeris-1 DecisionAPIClosed APIBase model: undisclosed71.0I 59C 88S 87K 60est.$0.021
- 26Sieve-31B-v2Usefine-tune · 31B · BF16Base model: google/gemma-4-31B-itsource70.7I 69C 93S 86K 51est.$0.043
- 27Torchcast Decision 4B (v1.1)Usefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource70.7I 54C 83S 94K 65est.$0.015
- 28Myles-G 31BAPIClosed APIBase model: google/gemma-4-31B-itmodel card*770.6I 74C 84S 86K 51est.$0.043
- 29Arcbright Elf 1 31BUsefine-tune · 31B · bf16Base model: google/gemma-4-31B-itsource70.5I 72C 92S 90K 50est.$0.047
- 30Jeff-1.1-MediumUsefine-tune · 12B · BF16Base model: google/gemma-4-12B-itsource70.5I 74C 89S 74K 54est.$0.035
- 31decider-12b v2Usefine-tune · 12BBase model: google/gemma-4-12B-itsource70.3I 62C 81S 90K 58est.$0.026
- 32Xevn-4BUsefine-tune · 4B · bf16Base model: Qwen/Qwen3.5-4Bsource70.2I 52C 85S 92K 66est.$0.014
- 33Jev 1.14.0APIJev referenceBase model: undisclosed70.2I 60C 90S 91K 55est.$0.032
- 34torchcast-decision-12bUseadapter · 12BBase model: google/gemma-4-12B-itsource69.8I 61C 83S 91K 56est.$0.028
- 35decider-12b v1, stock Gemma-4-12B-itUse12BBase model: google/gemma-4-12B-itsource69.8I 60C 83S 90K 58est.$0.026
- 36metask-jev-rain-31B-trainedUseadapter · 31B · W4A16 INT4Base model: google/gemma-4-31B-itsource69.6I 71C 87S 82K 51est.$0.044
- 37vLLM-SR d3-miniUsefine-tune · 4B · bf16Base model: Qwen/Qwen3.5-4Bsource69.5I 51C 87S 93K 64est.$0.016
- 38Bobcat 1.3Usefine-tune · 31B · fp8Base model: google/gemma-4-31B-itsource69.4I 68C 88S 89K 50est.$0.047
- 39Microsoft-Decision-1 (Azure Foundry)APIClosed APIBase model: Qwen3.5-9Bvendor statementvendor statement*869.1I 57C 84S 81K 61$0.019
- 40wity-1APIClosed APIBase model: undisclosedBase-model reference price (base undisclosed at the author's request): 43.0 (would be #91)69.0I 68C 86S 68K 59$0.024
- 41Chon-1.0-12B-FastUseadapter · 12B · bf16Base model: google/gemma-4-12B-itsource68.9I 60C 80S 84K 59est.$0.024
- 42Winnow-12B Q8Usemerge · 12B · GGUF Q8_0Base model: google/gemma-4-12B-itsource68.8I 60C 83S 86K 57est.$0.028
- 43deck-31BUse31B · FP8Base model: google/gemma-4-31B-itsource68.6I 73C 82S 86K 50est.$0.048
- 44Jeff-1.0-LargeUsefine-tune · 31BBase model: google/gemma-4-31B-itsource68.5I 72C 91S 88K 49est.$0.049
- 45CygnetUse12BBase model: google/gemma-4-12B-itsource68.5I 55C 87S 91K 56est.$0.028
- 46Wald 26B-A4BUsefine-tune · 26B-A4B · bf16Base model: google/gemma-4-26B-A4B-itsource68.2I 66C 88S 92K 50est.$0.048
- 47decisio v0.8.0 on gemma-4-12B-itUse12BBase model: google/gemma-4-12B-itsource68.1I 52C 89S 86K 59est.$0.023
- 48attest-31B-v1.0APIbase · 31BBase model: undisclosed67.9I 65C 74S 96K 51est.$0.044
- 49DOPEST 1.0 31BUsefine-tune · 31B · BF16Base model: google/gemma-4-31B-itsource67.8I 74C 90S 82K 49est.$0.049
- 50Sage 1.4APIClosed APIBase model: undisclosed67.7I 64C 83S 69K 59est.$0.022
- 51decisio v0.9.0 on gemma-4-12B-itUse12BBase model: google/gemma-4-12B-itsource67.7I 52C 88S 85K 59est.$0.023
- 52Jev-OmniUsefine-tuneBase model: google/gemma-4-12B-itsource67.6I 56C 87S 85K 56est.$0.029
- 53metask-jev-rain-12B-trainedUseadapter · 12B · W4A16 INT4Base model: google/gemma-4-12B-it-qat-q4_0-unquantizedsource67.3I 66C 87S 72K 53est.$0.038
- 54Bobcat Flash 1.2Usefine-tune · 26B / 4B active · FP8 on loadBase model: google/gemma-4-26B-A4B-itsource67.1I 66C 81S 91K 50est.$0.048
- 55Xor 26B-A4BUsemerge · 26B-A4BBase model: google/gemma-4-26B-A4B-itsource66.8I 57C 85S 91K 51est.$0.044
- 56SPX-CD FlashAPISystemBase model: Qwen/Qwen3.6-35B-A3Bmodel card*966.6I 58C 90S 79K 52est.$0.039
- 57Messier One v0.4Usefine-tune · 4BBase model: undisclosed66.1I 52C 73S 89K 61est.$0.020
- 58decider chat on Gemma-4-31B-itUse31BBase model: google/gemma-4-31B-itsource66.0I 59C 82S 86K 51est.$0.044
- 59Hopper 12B trainedUsefine-tune · 12BBase model: google/gemma-4-12B-itsource65.6I 52C 85S 84K 56est.$0.029
- 60metask-jev-rain-31B-trained service v3Useadapter · 31B · W4A16 INT4Base model: google/gemma-4-31B-itsource65.6I 73C 87S 73K 49est.$0.049
- 61Surogate Rune 26B-A4B v3Use26B-A4BBase model: google/gemma-4-26B-A4B-itsource*1065.0I 56C 91S 87K 50est.$0.048
- 62GEV-26B-DecideUsedistilled · 26B-A4BBase model: google/gemma-4-26B-A4B-itsource64.2I 50C 90S 90K 51est.$0.043
- 63SPX-CD-OmniUsefine-tune · 12BBase model: google/gemma-4-12B-itsource63.1I 49C 87S 89K 58est.$0.025
- 64Metask rain 4BUsefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource62.5I 51C 81S 78K 52est.$0.040
- 65shisa-de-1Usefine-tune · 26B-A4BBase model: google/gemma-4-26B-A4B-itsource62.4I 50C 81S 93K 50est.$0.047
- 66OpenAI DecisionsAPIClosed APIBase model: undisclosed62.4I 57C 90S 88K 49$0.052
- 67InstinctAPIClosed APIBase model: Qwen3.8-27Bsource*1162.2I 48C 93S 87K 64$0.015
- 68Blink v0.3 26B-A4BUsefine-tune · 26B-A4B · NVFP4Base model: google/gemma-4-26B-A4B-itsource61.1I 49C 91S 91K 50est.$0.048
- 69HS1 0.1.2APIClosed APIBase model: google/gemma-4-26B-A4B-it (self-reported)*1261.1I 53C 76S 94K 49est.$0.050
- 70baselabs elephant-1Usefine-tune · 9B · bf16Base model: Qwen/Qwen3.5-9Bsource60.9I 56C 85S 90K 48est.$0.052
- 71Diffusion JevUse26B / 4B activeBase model: google/diffusiongemma-26B-A4B-itsource59.3I 55C 61S 86K 50est.$0.048
- 72René-1 31B FP8Usefine-tune · 31B · FP8Base model: google/gemma-4-31B-itsource55.8I 62C 90S 87K 46est.$0.063
- 73Amazon Strands Decider 26B-A4BUseadapter · 26B-A4B · bf16Base model: google/gemma-4-26B-A4B-itsource54.5I 47C 87S 88K 50est.$0.045
- 74Wald 4B v2.1Usefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource54.3I 52C 87S 93K 46est.$0.062
- 75pplx-decider-flash-v1.2Usefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource54.0I 45C 78S 91K 64est.$0.015
- 76Bespoke Nimble 9B v3Useadapter · 9BBase model: Qwen/Qwen3.5-9Bsource53.5I 57C 77S 89K 46est.$0.063
- 77Aplomb 1Usefine-tune · 5.3BBase model: Qwen/Qwen3.5-4Bsource52.9I 45C 89S 89K 58est.$0.025
- 78Respan Span-02UseLLM decoderBase model: google/gemma-4-26B-A4B-itsource51.5I 61C 78S 86K 45est.$0.068
- 79Seb-9B-v2Usefine-tune · 9B · bf16Base model: Qwen/Qwen3.5-9Bsource50.0I 46C 84S 92K 49est.$0.051
- 80APUS-OpenJev-v1-9BUsefine-tune · 9BBase model: Qwen/Qwen3.5-9Bsource49.2I 47C 67S 82K 49est.$0.051
- 81Arcbright Elf 1 12BUsefine-tune · 12B · bf16Base model: google/gemma-4-12B-itsource48.8I 44C 86S 92K 57est.$0.026
- 82Jeffan-31b-1710Usefine-tune · 31B · bf16Base model: google/gemma-4-31B-itsource48.7I 72C 92S 85K 43est.$0.081
- 83Plumb-4BUsefine-tune · 4BBase model: alibiserikbay/JevK5source47.9I 43C 88S 94K 63est.$0.017
- 84SPX-CD ProAPISystemBase model: Qwen/Qwen3.8-27Bmodel card*1347.6I 65C 92S 77K 43est.$0.079
- 85Messier One v0.2Usefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource45.3I 42C 74S 93K 65$0.015
- 86spark-s1-4b-v6Useadapter · 4BBase model: Qwen3.5-4Bsource44.8I 43C 65S 86K 60est.$0.021
- 87KnowLine 31B-Gen1Usefine-tune · 31B · bf16Base model: google/gemma-4-31B-itsource44.4I 77C 93S 81K 41est.$0.093
- 88swanOneUseNVFP4Base model: Mia-AiLab/Qwen3.8-Flash-Next-NVFP4source43.8I 56C 89S 83K 42est.$0.085
- 89AJev lm6qUseadapter · 31B · bf16Base model: google/gemma-4-31B-itsource43.7I 71C 91S 82K 41est.$0.093
- 90Quyet-1.0-MediumUseadapter · 4BBase model: Qwen/Qwen3.5-4Bsource43.6I 42C 77S 89K 63est.$0.017
- 91Bespoke Nimble-v4APIClosed APIBase model: undisclosed43.2I 41C 87S 89K 61est.$0.020
- 92Turion Jev 31BUsefine-tune · 31B · fp8Base model: google/gemma-4-31B-itsource42.9I 71C 91S 86K 41est.$0.096
- 93Vansa-3.4APIClosed APIBase model: Qwen3.5-4B (self-reported)*1442.6I 41C 87S 95K 62$0.019
- 94JPT-4BUsemerge · 4BBase model: Qwen/Qwen3.5-4Bsource42.2I 41C 87S 88K 64est.$0.016
- 95levUseadapter · 4BBase model: Qwen/Qwen3.5-4Bsource42.1I 41C 82S 88K 62est.$0.019
- 96NInfer Qwen3.8-Flash-Next mixedUseLLM decoderBase model: Qwen3.8-Flash-Nextsource41.9I 49C 90S 89K 43est.$0.082
- 97decider-4b v2Useadapter · 4BBase model: Qwen/Qwen3.5-4B-Basesource41.1I 40C 90S 90K 65est.$0.015
- 98baselabs mammoth-1Usefine-tune · 27B · bf16Base model: Qwen/Qwen3.5-27Bsource41.1I 76C 94S 84K 40est.$0.104
- 99Jeffan-4B-2010Usefine-tune · 4B · bf16Base model: Qwen/Qwen3.5-4Bsource41.1I 40C 88S 93K 64est.$0.016
- 100jev-localUse9BBase model: Qwen/Qwen3.5-9Bsource41.0I 42C 73S 73K 59est.$0.024
- 101GPT-6 Luna (low reasoning effort)APIClosed APIBase model: undisclosed40.4I 96C 98S 72K 39$0.108
- 102Autoloops – Gemma 4 31B ITAPIClosed APIBase model: google/gemma-4-31B-itvendor statement*1540.2I 58C 87S 84K 40$0.096
- 103Wald 4BUsefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource39.7I 41C 78S 82K 54est.$0.033
- 104GPT-6 Luna (default medium reasoning effort)APIClosed APIBase model: undisclosed39.1I 97C 100S 71K 38$0.113
- 105ClassOne Qwen 3.5 9BUse9BBase model: Qwen/Qwen3.5-9Bsource38.5I 42C 59S 90K 49est.$0.049
- 106JevK5 v0.3UseadapterBase model: Qwen/Qwen3.5-4Bsource37.4I 39C 90S 94K 63est.$0.017
- 107metask-jev-4bUse4BBase model: Qwen3.5-4Bsource37.3I 39C 84S 90K 58est.$0.026
- 108deck-4B v1.0Useadapter · FP8Base model: alibiserikbay/JevK5 (built on Qwen/Qwen3.5-4B)source 1source 237.3I 39C 89S 90K 63est.$0.017
- 109GLiClass v4 9BnewUsefine-tune · 9B · BF16Base model: Qwen/Qwen3.5-9Bsource37.0I 40C 87S 86K 51est.$0.042
- 110Kahn1 4BUsefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource37.0I 40C 84S 89K 50est.$0.046
- 111Clef-FlashUsefine-tuneBase model: Qwen/Qwen3.5-9BsourceWorkers AI price scenario (latency unmeasured): 39.1 (would be #104)36.6I 42C 85S 88K 47est.$0.059
- 112Decision-4BUseadapterBase model: Qwen/Qwen3.5-4Bsource35.4I 38C 80S 91K 66est.$0.014
- 113decisor-4bUsefine-tune · 4B · FP8Base model: Qwen/Qwen3.5-4Bsource35.3I 40C 52S 92K 64est.$0.016
- 114janus 4BUsefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource35.2I 38C 88S 93K 64est.$0.015
- 115Qwen3.5-9B Jev-like data-mix v2Useadapter · 9BBase model: Qwen/Qwen3.5-9Bsource33.6I 42C 80S 85K 46est.$0.065
- 116Seb-9BUsefine-tune · 9B · bf16Base model: Qwen/Qwen3.5-9Bsource32.6I 38C 88S 91K 49est.$0.051
- 117JevK5 v0.2.0Useadapter · 4BBase model: Qwen3.5-4Bsource32.1I 36C 89S 91K 63est.$0.017
- 118Tern-35B-A3BUsefine-tune · 35B-A3B · bf16Base model: Qwen/Qwen3.6-35B-A3Bsource32.1I 44C 89S 91K 42est.$0.085
- 119Amos 1 mediumAPI9BBase model: Qwen/Qwen3.5-9Bmodel card*1631.6I 38C 88S 86K 49est.$0.051
- 120Aleria Judge 1.1.1Use31B · bf16Base model: undisclosed31.2I 68C 86S 86K 36est.$0.138
- 121JevOneUseadapterBase model: Qwen/Qwen3.6-35B-A3Bsource31.1I 46C 92S 90K 40est.$0.101
- 122Fastino GLiDEAPIClosed APIBase model: undisclosed30.6I 66C 88S 77K 36$0.137
- 123Jebadiah 9B v2.1Usemerge · 9BBase model: Qwen/Qwen3.5-9Bsource30.3I 37C 90S 91K 49est.$0.051
- 124Verdin-4BUsemerge · 4BBase model: Qwen/Qwen3.5-4Bsource29.9I 35C 84S 90K 62est.$0.018
- 125TypeCastLM 1.4.0Use3.8BBase model: Qwen/Qwen3.5-4Bsource29.8I 35C 77S 93K 65est.$0.015
- 126blink-4bUsemerge · 4BBase model: Qwen/Qwen3.5-4Bsource29.5I 35C 88S 91K 64est.$0.016
- 127jiwo 4bUsefine-tune · 4.2BBase model: Qwen/Qwen3.5-4Bsource28.9I 35C 86S 92K 64est.$0.016
- 128Imajev-4BUseadapter · 4BBase model: Qwen/Qwen3.5-4Bsource28.6I 35C 89S 90K 63est.$0.017
- 129SimpleJev Qwen3.6-35B-A3BAPI35B-A3BBase model: Qwen/Qwen3.6-35B-A3Bvendor statement*1727.5I 56C 80S 76K 35est.$0.145
- 130Uniphore-4B-DeciderUsefine-tune · 4BBase model: undisclosed27.2I 34C 80S 87K 65est.$0.015
- 131Decision 2.0 Lux 9BUseLLM decoderBase model: vllm-sr/Decision-1.0-Lux-9Bsource27.1I 36C 87S 90K 48est.$0.052
- 132HopperUseadapter · 4BBase model: Qwen/Qwen3.5-4Bsource26.5I 34C 89S 91K 62est.$0.018
- 133Malkuth-4BUseadapter · 4BBase model: Qwen/Qwen3.5-4B-Basesource26.2I 34C 88S 89K 56est.$0.030
- 134jqvUse32BBase model: Qwen3-32Bsource25.4I 34C 90S 82K 51est.$0.042
- 135JPT-35B-A3BUsemerge · 35B-A3BBase model: Qwen/Qwen3.5-35B-A3Bsource25.4I 51C 90S 90K 34est.$0.163
- 136decider-35b-a3bUsefine-tune · 35B-A3BBase model: Qwen/Qwen3.5-35B-A3B-Basesource24.8I 49C 84S 91K 34est.$0.154
- 137CoCo-Decision-4BUsefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource24.6I 33C 81S 91K 65est.$0.014
- 138Manchego v2.1Useadapter · 4BBase model: Qwen/Qwen3.5-4Bsource24.5I 33C 84S 91K 64est.$0.015
- 139Decision 4B v1.2Useadapter · 4BBase model: Qwen/Qwen3.5-4Bsource24.5I 32C 88S 94K 63est.$0.017
- 140Standard One 3B SH v4Usefine-tune · 3B · bf16Base model: StandardThinking/StandardOne-3Bsource24.4I 34C 78S 93K 49est.$0.052
- 141Fastino GLiDE 1.2Usefine-tune · 27B · bf16Base model: undisclosed22.0I 76C 93S 85K 31est.$0.205
- 142Raw Qwen3 4B Instruct 2507 direct logitsUseoriginal · 4BBase model: Qwen/Qwen3-4B-Instruct-2507 (unmodified)source21.8I 34C 35S 91K 63est.$0.017
- 143StartLux-Decision v1.1 27B InstantUsefine-tune · 27BBase model: undisclosed21.6I 73C 92S 90K 30est.$0.209
- 144Standard One 3B SHUsefine-tune · 3BBase model: StandardThinking/StandardOne-3Bsource21.6I 33C 78S 86K 49est.$0.052
- 145RSI-Jev v6.1-VL 27BUsemerge · 27BBase model: Qwen/Qwen3.8-27Bsource21.6I 62C 85S 85K 31est.$0.197
- 146GPT-5.6 LunaAPIClosed APIBase model: undisclosed21.4I 93C 94S 73K 30$0.213
- 147JEV-27BUsedistilled · 27BBase model: Qwen/Qwen3.8-27Bsource21.3I 57C 91S 86K 31est.$0.199
- 148torchcast-decision-27bUsemerge · 27BBase model: startlux-models/StartLux-Decision-27Bsource 1source 2*1821.1I 70C 90S 88K 30est.$0.210
- 149Bespoke Nimble 9BUseadapter · 9BBase model: Qwen3.5-9Bsource21.1I 43C 70S 86K 37est.$0.128
- 150Nimble-v4-maxUsefine-tune · 27BBase model: Qwen/Qwen3.8-27Bsource20.7I 71C 91S 85K 30est.$0.214
- 151pplx-decider-v1.2Usefine-tune · 27BBase model: Qwen/Qwen3.8-27Bsource20.6I 72C 90S 86K 30est.$0.215
- 152Perplexity Decider v1.1 27BUsefine-tune · 27BBase model: Qwen/Qwen3.8-27Bsource20.6I 76C 90S 86K 30est.$0.217
- 153vLLM-SR d3-turboUsefine-tune · 27BBase model: undisclosed20.5I 75C 90S 86K 30est.$0.218
- 154vLLM-SR d3Usefine-tune · 26BBase model: Qwen/Qwen3.8-27Bsource20.5I 73C 91S 87K 30est.$0.218
- 155reflex 4BUseadapter · 4BBase model: Qwen/Qwen3.5-4Bsource20.4I 31C 87S 69K 63est.$0.017
- 156Vera LargeUsefine-tune · 27B · bf16Base model: Qwen/Qwen3.8-27Bsource20.3I 68C 91S 88K 30est.$0.217
- 157Amos 1 largeAPI27BBase model: Qwen/Qwen3.8-27Bauthor self-reportmodel card*1920.3I 68C 88S 83K 30est.$0.215
- 158JADEUseadapter · 27BBase model: Qwen/Qwen3.8-27Bsource20.2I 48C 51S 85K 34est.$0.164
- 159Nagi-ENORMOUS 27BUseadapter · 27BBase model: Qwen/Qwen3.8-27Bsource19.8I 62C 85S 86K 30est.$0.216
- 160typecastlmUse4BBase model: Qwen3.5-4Bsource19.7I 30C 83S 93K 64est.$0.016
- 161Jebadiah 27BUsemerge · 27BBase model: Qwen/Qwen3.8-27Bsource19.2I 57C 83S 85K 30est.$0.217
- 162iota-decide-4bUsefine-tune · 4B · bf16Base model: Qwen/Qwen3.5-4B-Basesource18.9I 29C 85S 87K 63est.$0.017
- 163Decision 2.0 Vega 27BUseadapter · 27BBase model: Qwen/Qwen3.8-27Bsource 1source 218.9I 60C 82S 83K 30est.$0.222
- 164Jebadiah 9B v2Usemerge · 9BBase model: Qwen/Qwen3.5-9Bsource18.8I 31C 83S 91K 49est.$0.051
- 165Standard One 8B SHUsefine-tune · 8BBase model: StandardThinking/StandardOne-8Bsource18.7I 34C 85S 84K 43est.$0.078
- 166Decision 4B v1.1Useadapter · 4BBase model: Qwen/Qwen3.5-4Bsource18.7I 29C 85S 94K 63est.$0.017
- 167AutoJev-27BUsefine-tune · 27BBase model: Qwen/Qwen3.8-27Bsource18.5I 56C 88S 88K 29est.$0.226
- 168pjev-2bUsefine-tune · 2.5BBase model: openbmb/MiniCPM5-2Bsource18.1I 29C 84S 93K 64est.$0.016
- 169Eikos-27BUseadapter · 27BBase model: Qwen/Qwen3.8-27Bsource18.1I 64C 90S 88K 29est.$0.238
- 170Instinct Dual 4BAPIClosed APIBase model: Qwen3.5-4Bsource*2018.0I 28C 86S 89K 75$0.0068
- 171NInfer Qwen3.8-27B NVFP4Use27BBase model: Qwen3.8-27Bsource17.7I 52C 84S 90K 29est.$0.231
- 172Compass 0.2.0Useadapter · 4BBase model: Qwen/Qwen3.5-4Bsource17.6I 29C 83S 83K 58est.$0.026
- 173JEB-35B-A3BUsefine-tune · 35B-A3BBase model: Qwen/Qwen3.6-35B-A3Bsource17.3I 33C 90S 93K 43est.$0.080
- 174OpenJevUse26B-A4BBase model: google/diffusiongemma-26B-A4B-itsource 1source 217.3I 70C 81S 73K 29est.$0.241
- 175Open-Jev 9BUseadapter · 9BBase model: Qwen/Qwen3.5-9Bsource17.0I 44C 80S 74K 33est.$0.170
- 176Gemini 3.1 Flash-LiteAPIClosed APIBase model: Gemini 3 Pro (Google)model card*2116.9I 59C 61S 79K 29$0.230
- 177Standard One 8BUseadapter · 8BBase model: mistralai/Ministral-3-8B-Instruct-2512-BF16source16.9I 33C 84S 92K 43est.$0.078
- 178ClefUsefine-tuneBase model: Qwen/Qwen3.8-27BsourceWorkers AI price scenario (latency unmeasured): 29.4 (would be #127)16.7I 59C 92S 83K 28est.$0.249
- 179ArseneLupin v1.1Useadapter · 4BBase model: Qwen/Qwen3.5-4Bsource16.5I 28C 82S 90K 61est.$0.020
- 180system-one-openAPIadapter · 2BBase model: Gemma 4 E2Bsource16.4I 28C 76S 79K 68est.$0.011
- 181JEV Qwen3.5-9B Base NVFP4Use9B · NVFP4Base model: ig1/Qwen3.5-9B-NVFP4source16.0I 29C 85S 93K 48est.$0.056
- 182Arbiter-4BUsefine-tune · 4B · bf16Base model: Qwen/Qwen3.5-4Bsource15.9I 27C 77S 92K 66est.$0.013
- 183Decider-4BUsefine-tune · 4B · bf16Base model: Qwen/Qwen3.5-4Bsource15.7I 27C 87S 92K 66est.$0.013
- 184ClassOne Gemma 4 E4BUsefine-tune · E4BBase model: google/gemma-4-E4B-itsource15.6I 28C 49S 92K 69est.$0.011
- 185SemIfUse4BBase model: Qwen/Qwen3.5-4Bsource15.5I 27C 84S 90K 63est.$0.017
- 186openjev-sglangAPI35B-A3BBase model: Qwen/Qwen3.6-35B-A3Bsource15.4I 39C 79S 78K 36est.$0.140
- 187DecisionNode Edge 0.9.2Usefine-tune · 2B · bf16Base model: Qwen/Qwen3.5-2B-Basesource15.3I 27C 77S 93K 71est.$0.0091
- 188RYOTIDE-Qwen9Usepipeline · 9B · BF16Base model: Qwen/Qwen3.5-9Bsource14.7I 28C 77S 86K 49est.$0.049
- 189ClassOne Qwen 3.5 4BUsefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource14.3I 27C 49S 91K 65est.$0.015
- 190Jez-large-1.0.0Usefine-tune · E4B · bf16Base model: google/gemma-4-E4B-itsource13.8I 26C 83S 91K 71est.$0.0096
- 191RYOTIDE-QwenUsepipeline · 4BBase model: undisclosed12.9I 25C 87S 89K 62est.$0.018
- 192Chon-1.0-12BUsefine-tune · 12BBase model: google/gemma-4-12B-itsource12.8I 27C 59S 65K 49est.$0.050
- 193NomaUseLLM decoderBase model: Qwen/Qwen3.5-4B-Basesource12.7I 25C 84S 93K 62est.$0.019
- 194KnowLine 35B-A3B-Gen1.5Usefine-tune · 35B-A3B · fp8Base model: Qwen/Qwen3.5-35B-A3Bsource12.4I 60C 93S 92K 25est.$0.323
- 195Bev / Bonsai 27BUse27B · GGUF PQ2_0Base model: Qwen/Qwen3.8-27Bsource11.7I 44C 89S 73K 28est.$0.247
- 196LitJevUse27BBase model: Qwen/Qwen3.8-27Bsource11.5I 43C 90S 68K 28est.$0.244
- 197local-jev Qwen3.5-4BUsefine-tune · 4BBase model: Qwen/Qwen3.5-4Bsource11.4I 24C 86S 86K 59est.$0.023
- 198system-oneUse8BBase model: Qwen3-8Bsource11.0I 29C 30S 91K 45est.$0.068
- 199StartLux-Decision v1.1 27B AutoThinkUsefine-tune · 27BBase model: undisclosed10.8I 67C 91S 75K 24est.$0.354
- 200kev 8BUseadapter · 8BBase model: Qwen/Qwen3-8B-Basesource10.6I 29C 71S 88K 40est.$0.097
- 201Decision 2BUseadapter · 2BBase model: openbmb/MiniCPM5-2Bsource10.3I 23C 87S 90K 66est.$0.013
- 202OpenSourceJevUseadapter · 4B · GGUF Q4_K_MBase model: Qwen3.5-4Bsource10.2I 23C 74S 76K 69est.$0.011
- 203Imajev-9BUseadapter · 9BBase model: Qwen/Qwen3.5-9Bsource9.8I 37C 88S 79K 31est.$0.203
- 204open-alternative-jevUse4BBase model: Qwen3.5-4Bsource9.4I 22C 76S 91K 63est.$0.017
- 205Raw Qwen3 8B direct logitsUseoriginal · 8BBase model: Qwen/Qwen3-8B-Basesource9.3I 27C 34S 88K 46est.$0.065
- 206Liquid AI d1-3BUsefine-tune · 3BBase model: LiquidAI/LFM2.5-VL-3Bsource8.9I 22C 77S 94K 51est.$0.044
- 207Krino-2BUseadapter · 2BBase model: Qwen/Qwen3.5-2B-Basesource8.7I 22C 63S 91K 61est.$0.020
- 208classify-1APIClosed APIBase model: undisclosed8.5I 22C 83S 84K 51$0.044
- 209Imajev-2BUseadapter · 2BBase model: Qwen/Qwen3.5-2Bsource7.9I 21C 87S 81K 52est.$0.041
- 210decider-2bUseadapter · 2BBase model: Qwen/Qwen3.5-2B-Basesource7.7I 21C 69S 94K 65est.$0.015
- 211Nemotron Diffusion 8BUse8BBase model: nvidia/Nemotron-Labs-Diffusion-8Bsource7.3I 20C 75S 94K 55est.$0.030
- 212D1A on Gemma 4 E4BUsefine-tune · BF16Base model: google/gemma-4-E4Bsource7.1I 20C 82S 92K 61est.$0.021
- Nev-2B (re-measure pending under the 10 s rule)Usedistilled · 2B · GGUF Q8_0Base model: Qwen/Qwen3.5-2B-Basesource6.9I 20C 80S 55K 71est.$0.0092
- 213DeepSeek V4.1 FlashAPILLM decoderBase model: undisclosed6.9I 84C 95S 66K 20$0.480
- 214kev 4BUseadapter · 4BBase model: Qwen/Qwen3-4B-Basesource*226.6I 20C 68S 90K 66est.$0.014
- 215Clef-omniUsefine-tune · 30B-A3BBase model: Qwen/Qwen3-Omni-30B-A3B-Instructsource6.1I 34C 82S 87K 27est.$0.276
- 216Kev 27BUsefine-tune · 27BBase model: Qwen/Qwen3.8-27Bsource5.8I 64C 94S 83K 18est.$0.529
- 217FrontiersMind Lumma-fev-27BUsefine-tune · 27B · bf16Base model: Qwen/Qwen3.8-27Bsource5.3I 55C 89S 85K 18est.$0.544
- 218DecisionNode Edge Mini 0.9.2Usefine-tune · 0.8B · bf16Base model: Qwen/Qwen3.5-0.8B-Basesource5.3I 18C 78S 93K 80est.$0.0046
- 219jiwo 0.8bUsefine-tune · 0.75BBase model: Qwen/Qwen3.5-0.8Bsource4.1I 16C 79S 94K 78est.$0.0052
- 220Malkuth-2BUseadapter · 2BBase model: empero-ai/Qwen3.8-2B-Distillsource4.0I 16C 79S 92K 66est.$0.014
- 221Qwen3-Reranker-4BUseRerankerBase model: Qwen/Qwen3-4B-Basesource3.7I 16C 73S 81K 48est.$0.052
- 222SimpleJev Qwen3.8-27B (self-hosted, v1.6 pool)Use27BBase model: Qwen/Qwen3.8-27Bsource3.4I 58C 92S 76K 15est.$0.676
- 223ZeroEntropy zerank-2UseRerankerBase model: Qwen/Qwen3-4Bsource3.3I 16C 83S 82K 48est.$0.052
- 224EXAONE-4.0-1.2B-JEV v0.3Use1.2BBase model: LGAI-EXAONE/EXAONE-4.0-1.2Bsource3.3I 15C 81S 94K 58est.$0.025
- 225Open-Jev 2BUseadapter · 2BBase model: Qwen/Qwen3.5-2Bsource3.2I 22C 74S 77K 33est.$0.170
- 226SimpleJev Qwen3.8-27BAPI27BBase model: Qwen/Qwen3.8-27Bsource3.2I 53C 90S 76K 15est.$0.687
- 227Vega 4BUseadapter · 4BBase model: undisclosed3.1I 16C 57S 84K 47est.$0.058
- 228Open-Jev 27B v1.1Useadapter · 27BBase model: Qwen/Qwen3.8-27Bsource2.9I 52C 83S 72K 14est.$0.716
- 229Raw Phi-4 mini direct logitsUseoriginalBase model: microsoft/Phi-4-mini-instruct (unmodified)source2.5I 14C 57S 92K 53est.$0.036
- 230decision-machine-1APIClosed APIBase model: Qwen2.5-1.5B-Instruct (choices) + GLiNER2.5 (spans)vendor statement*232.4I 13C 79S 93K 58$0.024
- 231SimpleJevUse0.8BBase model: Qwen/Qwen3.5-0.8Bsource2.0I 13C 49S 59K 70est.$0.0098
- 232Jeetu Decider 2BUsefine-tune · 2B · bf16Base model: Qwen/Qwen3.5-2B-Basesource1.2I 10C 73S 91K 70est.$0.0098
- 233JevActAPIClosed APIBase model: Qwen3.5-2Bauthor self-report*241.1I 10C 47S 76K 69est.$0.011
- 234D1A on Gemma 4 E2BUsefine-tune · BF16Base model: google/gemma-4-E2Bsource0.8I 9C 82S 93K 61est.$0.021
- 235smalljev semantic-v9Useadapter · 2BBase model: openbmb/MiniCPM5-2B-Basesource0.8I 9C 77S 90K 61est.$0.020
- 236Gutsy 0.8B v0.3Usefine-tune · 0.8B · GGUF Q8_0Base model: Qwen/Qwen3.5-0.8Bsource0.8I 9C 76S 72K 78est.$0.0053
- 237Vega-U (Qwen3.5-4B reader)Useadapter · 4BBase model: undisclosed0.7I 8C 77S 89K 65est.$0.015
- 238PurpleMIST-Mini-1.0Usefine-tune · 2BBase model: Qwen/Qwen3.5-2B-Basesource0.7I 8C 78S 89K 71est.$0.0094
- 239kev 0.6BUseadapter · 0.6BBase model: Qwen/Qwen3-0.6B-Basesource0.5I 8C 66S 91K 80est.$0.0046
- 240jul fastUse2BBase model: openbmb/MiniCPM5-2Bsource0.5I 7C 78S 91K 69est.$0.011
- 241OpenDecisionUseEncoder / classifierBase model: MoritzLaurer/ModernBERT-large-zeroshot-v2.0source0.5I 7C 69S 86K 79est.$0.0050
- 242ClassOne Gemma 4 E2BUseE2BBase model: google/gemma-4-E2B-itsource0.5I 8C 23S 93K 70est.$0.0098
- Qwen3.5-0.8B Decision Model (re-measure pending under the 10 s rule)Usefine-tune · 0.8BBase model: Qwen/Qwen3.5-0.8B-Basesource0.5I 7C 78S 56K 80est.$0.0048
- 243Fastino GLiNER-2.5-DecideAPIfine-tune · 340MBase model: fastino/gliner2-large-v1source 1source 2*250.3I 7C 62S 85K 46$0.064
- 244Deem 0.8B v1UseLLM decoderBase model: Qwen/Qwen3.5-0.8Bsource0.3I 7C 30S 85K 80est.$0.0045
- 245Decision FastUseadapter · 0.6BBase model: Qwen/Qwen3-0.6B-Basesource0.3I 6C 75S 91K 80est.$0.0046
- 246Quyet-1.0-Small-ENUsefine-tune · 153MBase model: answerdotai/ModernBERT-basesource0.3I 6C 73S 95K 89est.$0.0023
- 247ClassOne Qwen 3.5 2BUsefine-tune · 2BBase model: Qwen/Qwen3.5-2Bsource0.2I 6C 33S 93K 70est.$0.0098
- 248Laya typed-decisionsUsefine-tune · 421MBase model: ModernBERT-largesource0.2I 5C 83S 69K 83est.$0.0038
- 249Vega-U (Qwen3.5-0.8B reader)Useadapter · 0.8BBase model: undisclosed0.2I 5C 67S 73K 79est.$0.0049
- 250WaterSheepUseEncoder / classifierBase model: answerdotai/ModernBERT-basesource0.2I 5C 66S 77K 69est.$0.011
- 251openJev Verdict 1.4UseEncoder / classifierBase model: knowledgator/gliclass-modern-base-v2.0source0.2I 5C 76S 81K 87est.$0.0028
- 252Raw Qwen3 0.6B direct logitsUseoriginal · 0.6BBase model: Qwen/Qwen3-0.6B-Basesource0.2I 6C 11S 93K 78est.$0.0056
- 253lev-350mUseadapterBase model: LiquidAI/LFM2.5-350Msource0.2I 5C 82S 94K 80est.$0.0046
- 254Raw Qwen3 1.7B direct logitsUseoriginal · 1.7BBase model: Qwen/Qwen3-1.7B-Basesource0.1I 5C 12S 93K 69est.$0.011
- 255kev 0.5BUseadapter · 0.5BBase model: Qwen/Qwen2.5-0.5Bsource0.1I 5C 61S 92K 80est.$0.0046
- 256Jez-small-1.0.0Usefine-tune · 0.8B · fp32Base model: Qwen/Qwen3.5-0.8Bsource0.1I 5C 82S 93K 80est.$0.0048
- 257openJev VerdictUsefine-tuneBase model: knowledgator/gliclass-modern-base-v2.0source 1source 20.1I 4C 49S 78K 87est.$0.0028
- 258Quyet-1.0-SmallUsefine-tune · 328MBase model: aisingapore/SEA-LION-ModernBERT-300Msource0.1I 4C 61S 95K 80est.$0.0048
- Reflex Instinct 0.6B (re-measure pending under the 10 s rule)Useadapter · 0.6BBase model: Qwen/Qwen3-0.6Bsource0.1I 4C 73S 59K 80est.$0.0047
- 259watt-flash-0.1Usefine-tune · 140MBase model: jhu-clsp/mmBERT-smallsource0.1I 4C 79S 74K 89$0.0023
- 260Quyet-1.0-TinyUsedistilled · 183MBase model: jhu-clsp/mmBERT-smallsource0.1I 4C 58S 95K 88est.$0.0024
- 261Phocinae-Largha-150M-v1Usefine-tune · 144MBase model: undisclosed0.1I 4C 57S 77K 95est.$0.0014
- 262Liquid AI d1-omni-600MUsefine-tune · 600MBase model: LiquidAI/LFM2.5-Encoder-350Msource0.0I 3C 62S 94K 73est.$0.0079
- 263Vega 0.8BUseadapter · 0.8BBase model: Qwen/Qwen3.5-0.8Bsource0.0I 3C 63S 89K 61est.$0.019
- 264open-jev-deberta-v3-largeUseadapterBase model: microsoft/deberta-v3-largesource0.0I 3C 68S 67K 78est.$0.0056
- modernbert-jev v2 (re-measure pending under the 10 s rule)Usefine-tune · 150M · ONNX INT8Base model: undisclosed0.0I 3C 67S 58K 70est.$0.0097
- 265Julia-1Usefine-tune · 144MBase model: jhu-clsp/mmBERT-smallsource0.0I 2C 17S 78K 90est.$0.0022
- 266verdict-smallUsefine-tune · 118MBase model: intfloat/multilingual-e5-smallsource0.0I 2C 73S 89K 100est.$0.0009
- 267Tacet SonataUsefine-tune · 144MBase model: jhu-clsp/mmBERT-smallsource0.0I 2C 46S 79K 89est.$0.0023
- 268LayaUsefine-tune · 421MBase model: ModernBERT-largesource0.0I 2C 63S 73K 85est.$0.0032
- 269Mixedbread mxbai-rerank-base-v2UseRerankerBase model: Qwen2.5 (size not stated)source 1source 20.0I 1C 87S 92K 60est.$0.021
- 270Laya multilingualUsefine-tune · 322MBase model: mmBERT-basesource0.0I 0C 34S 77K 82est.$0.0039
- 271Certo v1UseEncoder / classifierBase model: ModernBERT-largesource0.0I 0C 86S 92K 97est.$0.0013
- 272CLM-8BUsefine-tune · 8BBase model: Qwen/Qwen3-8Bsource0.0I 0C 40S 92K 51est.$0.045
- 273BAAI bge-reranker-v2-m3Usefine-tuneBase model: BAAI/bge-m3source0.0I 0C 88S 94K 59est.$0.023
- 274Alibaba GTE Reranker ModernBERT-baseUsefine-tuneBase model: answerdotai/ModernBERT-basesource0.0I 0C 83S 94K 69est.$0.011
- 275MirrorUseEncoder / classifierBase model: undisclosed0.0I 0C 52S 74K 89est.$0.0023
- 276Open Jev JSON CanvasUse26B-A4BBase model: google/diffusiongemma-26B-A4B-itsource0.0I 61C 0S 85K 49est.$0.049
- 277Qwen3.8 27BAPI27BBase model: Qwen/Qwen3.8-27Bmodel card*260.0I 33C 64S 55K 0est.$2.178
Base-model notes
- *1 Super Decider (base model: google/diffusiongemma-26B-A4B-it): Vendor model card; the hosted endpoint was not inspected.
- *2 Myles-G 12B (base model: google/gemma-4-12B-it): Public model-card metadata; weights gated, deployed endpoint not inspected.
- *3 Fastino GLiDE (base model: undisclosed): No public base model declared; proprietary vendor-hosted API.
- *4 Drex Max (base model: Qwen3.8-27B): Vendor's technical report. The vendor updated the model behind this API id after our 21 Oct measurement.
- *5 Bespoke Nimble-v4-max (base model: Qwen/Qwen3.8-27B): Vendor model card (published 18 Oct, after our 14 Oct measurement of the hosted API).
- *6 Super Decider (base model: google/diffusiongemma-26B-A4B-it): Vendor model card; the hosted endpoint was not inspected.
- *7 Myles-G 31B (base model: google/gemma-4-31B-it): Public model-card metadata; weights gated, deployed endpoint not inspected.
- *8 Microsoft-Decision-1 (base model: Qwen3.5-9B): Base model disclosed by Microsoft; the Decision-1 endpoint is evaluated as a hosted API offering.
- *9 SPX-CD Flash (base model: Qwen/Qwen3.6-35B-A3B): Vendor model card for the published adapter (9 Oct); the hosted Oct-4 checkpoint measured here is not stated to be identical.
- *10 Surogate Rune 26B-A4B v3 (base model: google/gemma-4-26B-A4B-it): This model also powers our System1 Models s1-pro service.
- *11 Instinct (base model: Qwen3.8-27B): Operator-reported; weights not publicly verifiable.
- *12 HS1 0.1.2 (base model: google/gemma-4-26B-A4B-it (self-reported)): Customer statement (order notes, 21 Oct 2026): a LoRA fine-tune of Gemma-4-26B-A4B-it served in FP8. Not publicly documented and not independently verified.
- *13 SPX-CD Pro (base model: Qwen/Qwen3.8-27B): Vendor model card for the published adapter (9 Oct); the hosted Oct-4 checkpoint measured here is not stated to be identical.
- *14 Vansa-3.4 (base model: Qwen3.5-4B (self-reported)): Developer-reported to Benchmark Heaven (private correspondence, 1 Oct 2026): built on Qwen3.5-4B via an open decision fine-tune, with Vansa's own adapters; the intermediate fine-tune is not named. Not publicly documented or independently verified.
- *15 Autoloops – Gemma 4 31B IT (base model: google/gemma-4-31B-it): Vendor page.
- *16 Amos 1 medium (base model: Qwen/Qwen3.5-9B): Model card; the hosted endpoint was not inspected.
- *17 SimpleJev Qwen3.6-35B-A3B (base model: Qwen/Qwen3.6-35B-A3B): Vendor README.
- *18 torchcast-decision-27b (base model: startlux-models/StartLux-Decision-27B): Direct base is StartLux-Decision-27B (startlux-models, CC-BY-NC-4.0), itself a modified Alibaba Qwen model; its card does not name the Qwen version. Qwen3.8-27B is inferred from the identical text_config.
- *19 Amos 1 large (base model: Qwen/Qwen3.8-27B): Vendor statement and model card; the hosted endpoint was not inspected.
- *20 Instinct Dual 4B (base model: Qwen3.5-4B): Operator-reported; weights not publicly verifiable.
- *21 Gemini 3.1 Flash-Lite (base model: Gemini 3 Pro (Google)): Google's own model.
- *22 kev 4B (base model: Qwen/Qwen3-4B-Base): Evaluated Qwen3 variant; later releases use a different base.
- *23 decision-machine-1 (base model: Qwen2.5-1.5B-Instruct (choices) + GLiNER2.5 (spans)): Vendor blog (19 Sep 2026); the vendor says the backend can change.
- *24 JevAct (base model: Qwen3.5-2B): Author's issue title only; not otherwise documented.
- *25 Fastino GLiNER-2.5-Decide (base model: fastino/gliner2-large-v1): Measured via Fastino's hosted API. A 340M English classifier (8k context), not built for multi-step reasoning: near chance on Noul and Score items; 23 very long items exceed its context. Our request follows Fastino's documented format. The open checkpoint scores 11.2 on the community Decision Index (rank 53 of 70; Jev 57.9). Fastino's larger GLiDE is a separate model, ranked separately on the API board.
- *26 Qwen3.8 27B (base model: Qwen/Qwen3.8-27B): The open model itself, not a fine-tune.
Adjust weights ↓
Weights are relative: each axis counts in proportion to its slider. The score stays a weighted harmonic mean with the low-axis gates; the gates still apply when an axis sits at 0. Only equal weights give the official JevBench Score and rank.
Score = 4 / (1/I + 1/C + 1/S + 1/K) (each 0–100; × (axis / 50)² for Intelligence, Speed or Cost below 50)
- Jev — reference (TypeSafe, closed) (2)
- Closed API (weights not public) (32)
- Open weights · LLM decoder (208)
- Open weights · diffusion LM (6)
- Open weights · encoder / classifier (21)
- Open weights · reranker (5)
- Base model control (no decision fine-tune, raw logits) (5)
- System (router / cascade / ensemble) (3)
- Striped bar = same system under the labelled alternative price assumption
- Shown, not ranked
Compare two systems
Pick any two. Radars for the score axes, capability by subject topic and use cases, chance-corrected competence per request type on the open and sealed sets, and competence per tier on each set. Further out is better on every spoke; the link keeps the pair.
- A: Jev 1.13.0REFERENCE · TypeSafe — Jev — reference (TypeSafe, closed) · Score 71.4 (#23)
- B: Super Decider — Open weights · diffusion LM · Score 78.5 (#1)
The four score axes
Capability by subject topic
Super Decider: No per-category values are published for this system.
Linear signed scale: −100 at the centre, 0 on the bold middle ring, 100 at the rim.
No per-category values are published for this pair.
What each category means · items per category
- Math & numbers — a calculation decides the answer: arithmetic, word problems, probability, dates, units. 235 items (174 open / 61 sealed)
- Coding & software — code, SQL, repositories, developer tools and IT systems. 92 items (67 open / 25 sealed)
- Rules, policy & law — applying written rules: company policies, contracts, regulations, eligibility. 548 items (260 open / 288 sealed)
- Finance & commerce — money: payments, refunds, invoices, orders, expenses, insurance payouts. 224 items (134 open / 90 sealed)
- Support & operations — support tickets, incidents, logistics, scheduling desks and routing work to a team. 282 items (131 open / 151 sealed)
- Everyday language — short everyday messages: intents, assistant requests, reading a detail out of a text. 115 items (98 open / 17 sealed)
- Safety & security — untrusted or injected instructions, fraud, moderation, access and security triage. 128 items (40 open / 88 sealed)
Use cases (TypeSafe categories)
Super Decider: No per-category values are published for this system.
Linear signed scale: −100 at the centre, 0 on the bold middle ring, 100 at the rim.
Linear signed scale: −100 at the centre, 0 on the bold middle ring, 100 at the rim.
Low sample, n < 30 — indicative only
| Category (items) | A: Jev 1.13.0 | B: Super Decider |
|---|---|---|
| Graphs and knowledge graphs (15) | 100.0 n=15 | — |
No published value for either system (under 15 answered items, or no per-category values): Scientific discovery (1), Semantic code linting (4), Feature extraction for predictive modeling (7), Recruiting (3), Moderation and trust and safety (12), Advertising (2).
Not drawn on the radar: with so few items a single answer moves a category score by several points, so these values are noise-prone. These values become spokes once the item pool reaches 30 items per category.
What each category means · items per category
- Search and retrieval — scoring query-to-candidate relevance, reranking results, selecting useful context for RAG. 54 items (18 open / 36 sealed)
- Model routing — deciding which LLM or model tier should handle a prompt; classifying a request's intent, domain, difficulty or risk to pick a model. 132 items (131 open / 1 sealed)
- LLM guardrails — checking LLM inputs, outputs and tool calls: prompt injection, jailbreaks, policy violations, tool-call errors, answer quality. 339 items (209 open / 130 sealed)
- Lead generation — matching companies or inbound messages to an ideal customer profile, buyer intent, prioritising and routing sales leads. 98 items (61 open / 37 sealed)
- Customer support — classifying tickets and customer intents, urgency, refunds, routing cases to a team, checking support replies. 389 items (176 open / 213 sealed)
- Insurance claims — first-notice-of-loss reports, claim complexity, missing information, claim triage. 31 items (14 open / 17 sealed)
- Financial crime — suspicious transactions, KYC, fraud alerts, entity matching for investigations. 31 items (10 open / 21 sealed)
- Legal and compliance — contracts, policies, regulations, eligibility rules, checking documents against written requirements. 508 items (261 open / 247 sealed)
- E-commerce marketplaces — product listings, product attributes, orders and deliveries, seller catalog policy. 71 items (48 open / 23 sealed)
- Risk assessment — estimating probabilities and severity of risks, incidents, vendor or underwriting risk. 78 items (39 open / 39 sealed)
Not drawn: Other — none of the above: general reasoning, arithmetic or date puzzles, everyday assistant requests without a business workflow. 159 items (76 open / 83 sealed) — not a use case of its own, so it is counted but not drawn.
No items in this release: Gaming, Demand forecasting.
Competence per request type, open / sealed
Linear signed scale: −100 at the centre, 0 on the bold middle ring, 100 at the rim.
Competence per tier — open set
Linear signed scale: −100 at the centre, 0 on the bold middle ring, 100 at the rim.
Competence per tier — sealed set
Linear signed scale: −100 at the centre, 0 on the bold middle ring, 100 at the rim.
How the categories were made
Chance-corrected competence (0 = chance, 100 = perfect) over the category's items, open and sealed pooled, all tiers: per request type the method's competence cell (Choice and Noul accuracy above chance, Score 1 - error / chance error), then averaged over types weighted by item count. Not part of the JevBench Score; compare systems within a category, not categories with each other. Values can be negative (below chance).
Every one of the 1,624 decisions was labelled with one subject topic (the v1.2 topic list, unchanged) and one or two TypeSafe use-case categories (docs.typesafe.ai/concepts/use-case-map, plus Other) by Winnow-12B Q8 (the model behind System1 Models s1-pro) on our own GPU pod; 1,118 of the 3,248 calls were first made on s1-pro itself and agree 98.2 % with the pod run. Sealed items stayed on our own infrastructure. About 5 % of public items were checked by hand (Claude Opus 5.5); the item-group rules below fix the systematic misses found.
- routing, routing_hard and tool_selection items (choose the model, agent or tool that handles a request) are Model routing first; the labeller tagged many by the request subject.
- tool_guardrail, unsafe_action, adequacy, judge_hard and policy_compliance items (check an agent tool call or a drafted answer before it goes out) are LLM guardrails first.
- lead_qualification items take the topic Finance & commerce (sales); the labeller read the scoring rubric as rules & law.
- A second use case is kept when the labeller gave it probability >= 0.15 and it is not Other.
All values as a table
| Spoke | A: Jev 1.13.0 | B: Super Decider |
|---|---|---|
| The four score axes | ||
| Intelligence | 63.6 | 78.9 |
| Calibration | 90.6 | 90.8 |
| Speed | 91.2 | 85.3 |
| Cost | 54.7 | 64.3 |
| Capability by subject topic | ||
| Math & numbers | 40.5 | No per-category values are published for this system. |
| Coding & software | 80.3 | No per-category values are published for this system. |
| Rules, policy & law | 73.7 | No per-category values are published for this system. |
| Finance & commerce | 83.4 | No per-category values are published for this system. |
| Support & operations | 91.5 | No per-category values are published for this system. |
| Everyday language | 96.1 | No per-category values are published for this system. |
| Safety & security | 67.7 | No per-category values are published for this system. |
| Use cases (TypeSafe categories) | ||
| Search & retrieval | 95.9 | No per-category values are published for this system. |
| Model routing | 96.5 | No per-category values are published for this system. |
| LLM guardrails | 67.0 | No per-category values are published for this system. |
| Lead generation | 88.1 | No per-category values are published for this system. |
| Customer support | 91.9 | No per-category values are published for this system. |
| Insurance claims | 67.5 | No per-category values are published for this system. |
| Financial crime | 58.3 | No per-category values are published for this system. |
| Legal & compliance | 74.0 | No per-category values are published for this system. |
| E-commerce | 89.8 | No per-category values are published for this system. |
| Risk assessment | 80.1 | No per-category values are published for this system. |
| Competence per request type, open / sealed | ||
| Choice · open | 79.0 | 82.9 |
| Choice · sealed | 75.8 | 81.3 |
| Noul · open | 54.5 | 84.0 |
| Noul · sealed | 46.1 | 84.0 |
| Score · open | 63.3 | 69.5 |
| Score · sealed | 63.0 | 71.6 |
| Competence per tier — open set | ||
| Easy | 93.9 | 97.7 |
| Standard | 63.2 | 77.2 |
| Judge | 67.7 | 75.0 |
| Hard | 65.6 | 79.5 |
| Competence per tier — sealed set | ||
| Easy | 78.4 | 88.6 |
| Standard | 68.7 | 86.2 |
| Judge | 66.9 | 80.3 |
| Hard | 58.9 | 72.7 |
Category radars count each answered item once from the pools named under each radar. API overlay rows use A4+P+L1+L2+L3; A5+P+L1+L2+L3; S+P+L1+L2+L3. Raw and unequated; cells under 15 answered items are omitted. Per-type and tier radars retain each row's original measurement pools: A4/A5 rows have 300 open plus 300 sealed items; full-set rows have S 1,200 plus P 300.
Languages
Raw chance-corrected competence per item language (0 = chance, 100 = perfect; can be negative), from each system's own measured items: Self-hosted rows use S 1,200 + P 300 + L1 (354 items) + L2 (33 items) + L3 (1,118 items), where answered. API rows re-run on A4/A5 use their 300/300-item sealed subset instead of S. Every row’s tag lists the pools it actually answered. L3 is a sealed language supplement (1,118 items), drawn 2026-10-07. Every language has at least 60 items in P ∪ L1 ∪ L2 ∪ L3. Header counts show the full S + P + L1 + L2 + L3 pool. Headline scores are unchanged; these raw cells are unequated, scored for language/category views only and outside the Composite. L3 items were written natively by Claude Sonnet 5.5, each solved blind and language-checked by GPT-6.1 Sol, with gold kept only when both agree or a second review confirms; no gold comes from Jev or any measured API. L3 is API-facing by design and is excluded from future headline draws. Rows with unfinished runs retain their actual coverage tags. L3 items were reviewed by an OpenAI model, so the L3-based language and category cells of OpenAI rows (OpenAI Decisions, GPT-6 Luna, GPT-5.6 Luna) carry that exposure; headline scores do not use L3. C1 adds English items for thin radar categories (everyday language, safety and the other use case). L4 adds 20 new Hindi items for Fastino GLiNER 2.5 Decide only. They were authored with OpenAI and independently solved and language-reviewed with Anthropic before sealing. Seventeen L4 requests succeeded; three HTTP 400 failures remain in the scored observations and do not count as completed coverage. L4 is outside headline and category scores and the common header counts. L5 adds 60 new items for OpenJev DeBERTa v3 Large only: 20 each in Arabic, Hindi and Greek. They were authored with OpenAI and independently solved and language-reviewed with Anthropic before sealing, then checked against the original 512-token native context before keeper selection. All 60 native requests succeeded. Previous failures remain scored and do not count as completed coverage. L5 is outside headline and category scores and the common header counts. Cells under 15 completed supported responses are left empty; recorded input refusals count toward coverage. A dagger (†) marks fewer than 30 completed responses. Competence retains all supported scored observations, including failures and refusals. The mixed-language group is listed first, then English (1,306 items); the other 21 languages and the mixed-language group share 1699 items.
313 systems
| 78 | 1306 | 95 | 90 | 88 | 84 | 83 | 83 | 80 | 80 | 77 | 75 | 75 | 75 | 73 | 73 | 72 | 72 | 71 | 71 | 70 | 69 | 65 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Super DeciderAPI · not measured | — | 81 | 87 | 77† | 76 | 86 | — | 76† | — | 80† | — | — | — | — | — | — | — | 72† | 81† | — | — | — | 66† |
| Myles-G 12BAPI · S+P (draw v1.6-regular-20261012) | — | 56 | 66 | 73† | 59 | 86 | — | 58† | — | 48† | — | — | — | — | — | — | — | 82† | 4† | — | — | — | 43† |
| kettei-fatAPI · S+P (draw v1.6-regular-20261012) | — | 80 | 90 | 76† | 86 | 87 | — | 81† | — | 81† | — | — | — | — | — | — | — | 79† | 86† | — | — | — | 78† |
| Fastino GLiDE (hosted API, 22 Oct 2026)API · S+P (draw v1.6-regular-20261012) | — | 76 | 74 | 78† | 79 | 93 | — | 66† | — | 47† | — | — | — | — | — | — | — | 70† | 49† | — | — | — | 94† |
| Torchcast Decision 4B (v1.3)self-hosted · S+P (draw v1.6-regular-20261012) | — | 66 | 72 | 75† | 71 | 68 | — | 82† | — | 62† | — | — | — | — | — | — | — | 70† | 27† | — | — | — | 66† |
| d1-ultraAPI · S+P (draw v1.6-regular-20261012) | — | 70 | 66 | 86† | 67 | 89 | — | 63† | — | 24† | — | — | — | — | — | — | — | 69† | 28† | — | — | — | 44† |
| SubQ Snap 1.1API · S+P (draw v1.6-regular-20261012) | — | 75 | 75 | 76† | 83 | 91 | — | 61† | — | 39† | — | — | — | — | — | — | — | 47† | 77† | — | — | — | 78† |
| Drex MaxAPI · S+P (draw v1.6-regular-20261012) | — | 79 | 72 | 85† | 82 | 88 | — | 76† | — | 54† | — | — | — | — | — | — | — | 71† | 70† | — | — | — | 94† |
| Bespoke Nimble-v4-maxAPI · S+P (draw v1.6-regular-20261012) | — | 71 | 70 | 65† | 73 | 82 | — | 71† | — | 27† | — | — | — | — | — | — | — | 34† | 59† | — | — | — | 61† |
| Super Decider (thinking)API · not measured | — | 82 | 81 | 80† | 67 | 88 | — | 74† | — | 78† | — | — | — | — | — | — | — | 68† | 81† | — | — | — | 83† |
| SubQ SnapAPI · S+P (draw v1.6-regular-20261012) | — | 74 | 68 | 62† | 72 | 84 | — | 67† | — | 39† | — | — | — | — | — | — | — | 51† | 61† | — | — | — | 66† |
| H2O-Lightning-4B v1.1self-hosted · S+P+L1+L2+L3 | 85 | 63 | 55 | 59 | 60 | 54 | 49 | 38 | 54 | 58 | 41 | 37 | 34 | 46 | 26 | 49 | 46 | 39 | 48 | 55 | 47 | 31 | 23 |
| Sage 1.3.0API · S+P+L1+L2+L3 | 80 | 71 | 48 | 55 | 42 | 49 | 44 | 45 | 57 | 51 | 36 | 44 | 43 | 32 | 26 | 44 | 51 | 65 | 62 | 50 | 59 | 37 | 43 |
| kettei-chubAPI · S+P (draw v1.6-regular-20261012) | — | 62 | 77 | 63† | 61 | 83 | — | 77† | — | 49† | — | — | — | — | — | — | — | 24† | 28† | — | — | — | 72† |
| H2O-Lightning-31B v1.0.7self-hosted · S+P (draw v1.6-regular-20261012) | — | 81 | 66 | 75† | 83 | 91 | — | 78† | — | 65† | — | — | — | — | — | — | — | 75† | 28† | — | — | — | 77† |
| Torchcast Decision 4B (v1.2)self-hosted · S+P (draw v1.6-regular-20261012) | — | 60 | 65 | 68† | 76 | 78 | — | 71† | — | 63† | — | — | — | — | — | — | — | 39† | 9† | — | — | — | 74† |
| d1API · S+P+L1+L2+L3 | 80 | 66 | 32 | 55 | 34 | 44 | 33 | 36 | 64 | 40 | 28 | 30 | 41 | 28 | 14 | 33 | 27 | 48 | 51 | 41 | 47 | 12 | 32 |
| Mercury DecideAPI · not measured | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| Decisio 12B tuned commitself-hosted · not measured | — | 60 | 70 | 66† | 61 | 83 | — | 55† | — | 52† | — | — | — | — | — | — | — | 56† | 39† | — | — | — | 71† |
| Chon-1.0-E4Bself-hosted · not measured | — | 50 | 68 | 75† | 68 | 83 | — | 71† | — | 45† | — | — | — | — | — | — | — | 61† | 25† | — | — | — | 52† |
| decisio v0.8.0 on gemma-4-31B-itself-hosted · not measured | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| basal-omni-1.0-ultraself-hosted · S+P (draw v1.6-regular-20261012) | — | 77 | 76 | 70† | 65 | 86 | — | 62† | — | 58† | — | — | — | — | — | — | — | 71† | 39† | — | — | — | 78† |
| Jev 1.13.0API · S+P+L1+L2+L3 | 76 | 68 | 38 | 44 | 22 | 48 | 35 | 34 | 36 | 36 | 30 | 15 | 43 | 11 | 7 | 34 | 20 | 36 | 43 | 23 | 47 | 19 | 21 |
| Microsoft-Decision-1 (Azure Foundry) — A2 re-runAPI · A2+P (draw v1.6-regular-20261012) | — | 59 | — | — | 53† | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| Quyet-1.0-Largeself-hosted · S+P+L1+L2+L3 | 84 | 76 | 47 | 57 | 56 | 66 | 54 | 40 | 66 | 55 | 43 | 52 | 59 | 43 | 29 | 64 | 59 | 67 | 67 | 46 | 55 | 45 | 64 |
| Celeris-1 DecisionAPI · S+P (draw v1.6-regular-20261012) | — | 59 | 59 | 51† | 56 | 88 | — | 72† | — | 43† | — | — | — | — | — | — | — | 52† | 11† | — | — | — | 55† |
| Sieve-31B-v2self-hosted · not measured | — | 71 | 67 | 51† | 67 | 91 | — | 51† | — | 48† | — | — | — | — | — | — | — | 40† | 44† | — | — | — | 61† |
| Torchcast Decision 4B (v1.1)self-hosted · S+P (draw v1.6-regular-20261012) | — | 54 | 62 | 65† | 72 | 61 | — | 80† | — | 46† | — | — | — | — | — | — | — | 33† | 0† | — | — | — | 59† |
| Myles-G 31BAPI · S+P (draw v1.6-regular-20261012) | — | 75 | 71 | 71† | 68 | 89 | — | 59† | — | 40† | — | — | — | — | — | — | — | 76† | 38† | — | — | — | 67† |
| Arcbright Elf 1 31Bself-hosted · not measured | — | 72 | 67 | 71† | 74 | 86 | — | 59† | — | 50† | — | — | — | — | — | — | — | 73† | 24† | — | — | — | 77† |
| Jeff-1.1-Mediumself-hosted · not measured | — | 75 | 82 | 76† | 88 | 94 | — | 85† | — | 29† | — | — | — | — | — | — | — | 88† | 52† | — | — | — | 78† |
| decider-12b v2self-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 87 | 65 | 51 | 58 | 61 | 40 | 49 | 55 | 61 | 49 | 54 | 43 | 55 | 46 | 43 | 57 | 59 | 67 | 58 | 56 | 40 | 58 | 36 |
| Xevn-4Bself-hosted · not measured | — | 49 | 63 | 72† | 44 | 72 | — | 77† | — | 56† | — | — | — | — | — | — | — | 37† | 25† | — | — | — | 70† |
| Jev 1.14.0API · S+P (draw v1.6-regular-20261012) | — | 63 | 73 | 73† | 42 | 75 | — | 51† | — | 0† | — | — | — | — | — | — | — | 39† | 0† | — | — | — | 44† |
| torchcast-decision-12bself-hosted · S+P+L1+L2+L3 | 86 | 65 | 49 | 64 | 47 | 48 | 39 | 53 | 64 | 50 | 44 | 47 | 45 | 44 | 33 | 54 | 40 | 68 | 53 | 54 | 44 | 58 | 44 |
| decider-12b v1, stock Gemma-4-12B-itself-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 85 | 62 | 48 | 50 | 49 | 42 | 46 | 54 | 61 | 51 | 42 | 46 | 47 | 38 | 32 | 49 | 44 | 61 | 50 | 49 | 40 | 51 | 43 |
| metask-jev-rain-31B-trainedself-hosted · S+P (draw v1.6-regular-20261012) | — | 71 | 69 | 68† | 63 | 90 | — | 57† | — | 46† | — | — | — | — | — | — | — | 73† | 36† | — | — | — | 78† |
| vLLM-SR d3-miniself-hosted · not measured | — | 52 | 62 | 57† | 56 | 73 | — | 80† | — | 36† | — | — | — | — | — | — | — | 30† | 0† | — | — | — | 28† |
| Bobcat 1.3self-hosted · not measured | — | 69 | 71 | 70† | 63 | 87 | — | 62† | — | 46† | — | — | — | — | — | — | — | 74† | 37† | — | — | — | 66† |
| Microsoft-Decision-1API · not measured | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| wity-1 (Wity, reasoning always)API · S+P+L1+L2+L3 | 75 | 73 | 57 | 68 | 32 | 50 | 33 | 44 | 46 | 43 | 29 | 40 | 17 | 36 | 32 | 29 | 41 | 44 | 55 | 43 | 31 | 18 | 5 |
| wity-1 (Wity, reasoning auto)API · S+P+L1+L2+L3 | 83 | 73 | 59 | 63 | 31 | 60 | 29 | 40 | 33 | 38 | 44 | 33 | 39 | 54 | 24 | 16 | 42 | 56 | 39 | 36 | 45 | 15 | 16 |
| Chon-1.0-12B-Fastself-hosted · not measured | — | 56 | 71 | 75† | 78 | 88 | — | 49† | — | 49† | — | — | — | — | — | — | — | 71† | 31† | — | — | — | 44† |
| Winnow-12B Q8self-hosted · S+P+L1+L2+L3 | 78 | 64 | 40 | 59 | 39 | 52 | 39 | 54 | 53 | 45 | 41 | 29 | 44 | 33 | 41 | 46 | 36 | 57 | 55 | 34 | 38 | 44 | 40 |
| deck-31Bself-hosted · S+P+L1+L2+L3 | 91 | 78 | 60 | 69 | 55 | 68 | 65 | 60 | 74 | 71 | 56 | 58 | 71 | 50 | 53 | 68 | 60 | 70 | 61 | 59 | 55 | 45 | 60 |
| Jeff-1.0-Largeself-hosted · S+P (draw v1.6-fastlane-20261009) | 99† | 71 | 89 | 67† | 64 | 91 | 56† | — | 98† | 49† | — | — | — | — | — | — | 77† | 84† | 62† | — | — | — | 59† |
| Cygnetself-hosted · S+P+L1+L2+L3 | 79 | 57 | 38 | 56 | 41 | 41 | 32 | 39 | 63 | 45 | 34 | 27 | 42 | 36 | 21 | 42 | 45 | 53 | 54 | 34 | 39 | 46 | 40 |
| Wald 26B-A4Bself-hosted · not measured | — | 64 | 65 | 68† | 74 | 72 | — | 60† | — | 36† | — | — | — | — | — | — | — | 58† | 29† | — | — | — | 78† |
| decisio v0.8.0 on gemma-4-12B-itself-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 79 | 56 | 39 | 55 | 39 | 42 | 33 | 38 | 63 | 41 | 29 | 25 | 34 | 35 | 25 | 37 | 44 | 48 | 56 | 38 | 40 | 42 | 27 |
| attest-31B-v1.0API · S+P (draw v1.6-regular-20261012) | — | 67 | 67 | 74† | 58 | 89 | — | 53† | — | 45† | — | — | — | — | — | — | — | 82† | 26† | — | — | — | 67† |
| DOPEST 1.0 31Bself-hosted · not measured | — | 74 | 60 | 75† | 69 | 91 | — | 66† | — | 51† | — | — | — | — | — | — | — | 79† | 42† | — | — | — | 78† |
| Sage 1.4API · S+P (draw v1.6-regular-20261012) | — | 65 | 65 | 73† | 65 | 83 | — | 67† | — | 39† | — | — | — | — | — | — | — | 34† | 52† | — | — | — | 61† |
| decisio v0.9.0 on gemma-4-12B-itself-hosted · S+P+L1+L2+L3 | 81 | 56 | 41 | 55 | 38 | 40 | 31 | 41 | 62 | 41 | 28 | 29 | 36 | 34 | 25 | 37 | 44 | 48 | 56 | 37 | 40 | 43 | 31 |
| Jev-Omniself-hosted · S+P+L1+L2+L3 | 86 | 56 | 38 | 58 | 59 | 44 | 45 | 37 | 58 | 50 | 38 | 27 | 49 | 28 | 25 | 50 | 38 | 57 | 55 | 47 | 48 | 46 | 37 |
| metask-jev-rain-12B-trainedself-hosted · S+P (draw v1.6-regular-20261012) | — | 64 | 74 | 81† | 72 | 85 | — | 49† | — | 45† | — | — | — | — | — | — | — | 59† | 25† | — | — | — | 77† |
| Bobcat Flash 1.2self-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 82 | 71 | 45 | 47 | 38 | 54 | 33 | 33 | 51 | 41 | 43 | 37 | 39 | 28 | 26 | 47 | 37 | 48 | 50 | 44 | 58 | 29 | 27 |
| Xor 26B-A4Bself-hosted · S+P+L1+L2+L3 | 80 | 61 | 46 | 40 | 48 | 56 | 51 | 36 | 52 | 44 | 34 | 45 | 43 | 32 | 33 | 50 | 40 | 68 | 50 | 50 | 50 | 42 | 46 |
| SPX-CD FlashAPI · S+P+L1+L2+L3 | 78 | 63 | 38 | 43 | 36 | 40 | 27 | 34 | 49 | 48 | 31 | 22 | 30 | 17 | 8 | 38 | 24 | 37 | 51 | 46 | 50 | 12 | 29 |
| Messier One v0.4self-hosted · S+P (draw v1.6-regular-20261012) | — | 51 | 68 | 62† | 70 | 72 | — | 65† | — | 39† | — | — | — | — | — | — | — | 32† | 18† | — | — | — | 55† |
| decider chat on Gemma-4-31B-itself-hosted · S+P+L1+L2+L3 | 75 | 64 | 58 | 54 | 34 | 53 | 50 | 45 | 56 | 59 | 45 | 39 | 52 | 45 | 32 | 43 | 44 | 51 | 44 | 43 | 44 | 44 | 49 |
| Hopper 12B trainedself-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 79 | 54 | 50 | 51 | 42 | 48 | 33 | 39 | 54 | 42 | 32 | 23 | 38 | 28 | 16 | 36 | 32 | 48 | 47 | 34 | 32 | 24 | 34 |
| metask-jev-rain-31B-trained service v3self-hosted · S+P (draw v1.6-regular-20261012) | — | 74 | 68 | 74† | 63 | 82 | — | 55† | — | 65† | — | — | — | — | — | — | — | 80† | 43† | — | — | — | 78† |
| Surogate Rune 26B-A4B v3 (v1.6 pool)self-hosted · S+P+L1+L2+L3 | 75 | 59 | 47 | 43 | 36 | 47 | 37 | 37 | 45 | 44 | 33 | 37 | 45 | 38 | 19 | 52 | 37 | 55 | 43 | 40 | 50 | 23 | 33 |
| GEV-26B-Decideself-hosted · S+P+L1+L2+L3 | 78 | 52 | 33 | 37 | 31 | 42 | 37 | 42 | 37 | 28 | 31 | 23 | 30 | 28 | 21 | 35 | 29 | 42 | 21 | 40 | 28 | 27 | 25 |
| SPX-CD-Omniself-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 80 | 53 | 35 | 40 | 46 | 36 | 29 | 32 | 50 | 44 | 41 | 34 | 44 | 26 | 31 | 48 | 22 | 44 | 56 | 35 | 38 | 24 | 34 |
| Metask rain 4Bself-hosted · S+P (draw v1.6-fastlane-20261009) | 100† | 48 | 65 | 74† | 60 | 62 | 61† | — | 30† | 8† | — | — | — | — | — | — | 0† | 41† | 0† | — | — | — | 73† |
| shisa-de-1self-hosted · S+P (draw v1.6-regular-20261012) | — | 45 | 61 | 76† | 45 | 88 | — | 52† | — | 30† | — | — | — | — | — | — | — | 52† | 25† | — | — | — | 50† |
| OpenAI DecisionsAPI · A5+P+L1+L2+L3L3 items were reviewed by an OpenAI model, so the L3-based language and category cells of this row carry that exposure; headline scores do not use L3. | 76 | 59 | 36 | 48 | 41 | 55 | 27 | 37 | 43 | 37 | 24 | 23 | 46 | 37 | 30 | 33 | 27 | 40 | 33 | 39 | 37 | 40 | 25 |
| InstinctAPI · A4+P+L1+L2+L3 | 67 | 58 | 22 | 35 | 32 | 28 | 27 | 31 | 52 | 48 | 20 | 18 | 34 | 19 | 23 | 22 | 39 | 39 | 39 | 42 | 43 | 13 | 7 |
| Blink v0.3 26B-A4Bself-hosted · S+P (draw v1.6-regular-20261010-a2) | — | 46 | 54† | 90† | 45 | 70 | 54† | 77† | 39† | 25 | — | — | — | — | — | — | — | 39 | — | — | — | — | 0† |
| HS1 0.1.2API · S+P (draw v1.6-regular-20261012) | — | 50 | 54 | 73† | 46 | 78 | — | 61† | — | 35† | — | — | — | — | — | — | — | 58† | 37† | — | — | — | 39† |
| baselabs elephant-1self-hosted · not measured | — | 53 | 74 | 72† | 62 | 72 | — | 54† | — | 32† | — | — | — | — | — | — | — | 70† | 24† | — | — | — | 55† |
| Diffusion Jevself-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 82 | 59 | 51 | 57 | 49 | 65 | 37 | 41 | 52 | 44 | 48 | 29 | 43 | 45 | 36 | 49 | 44 | 63 | 50 | 50 | 51 | 38 | 16 |
| René-1 31B FP8self-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 75 | 67 | 45 | 39 | 42 | 46 | 40 | 31 | 55 | 37 | 35 | 37 | 54 | 13 | 25 | 44 | 33 | 47 | 43 | 42 | 46 | 20 | 43 |
| Amazon Strands Decider 26B-A4Bself-hosted · not measured | — | 43 | 68 | 49† | 49 | 69 | — | 59† | — | 56† | — | — | — | — | — | — | — | 32† | 15† | — | — | — | 40† |
| Wald 4B v2.1self-hosted · S+P (draw v1.6-regular-20261010-a2) | — | 50 | 72† | 75† | 34 | 55 | 43† | 54† | 54† | 49 | — | — | — | — | — | — | — | 25 | — | — | — | — | 12† |
| pplx-decider-flash-v1.2self-hosted · S+P (draw v1.6-regular-20261012) | — | 47 | 66 | 65† | 54 | 74 | — | 86† | — | 24† | — | — | — | — | — | — | — | 0† | 0† | — | — | — | 22† |
| Bespoke Nimble 9B v3self-hosted · S+P+L1+L2+L3 | 80 | 61 | 45 | 50 | 35 | 44 | 30 | 30 | 56 | 41 | 27 | 7 | 40 | 32 | 17 | 36 | 23 | 32 | 42 | 60 | 54 | 20 | 37 |
| Aplomb 1self-hosted · S+P+L1+L2+L3The headline keeps the original published measurement. Its originally pinned Hugging Face revision (f8e2dc81) was later deleted upstream, so the language and category supplements (L1, L2, L3) were measured on 12 Oct 2026 with the current upstream revision d6743291 (all file hashes recorded) on an isolated pod. That revision reproduces the original answers: 98 % agreement on the 300 public items (published bar 95 %) and 97.5 % on all 1,500 headline items. Each supplement carries the original P300 answers by stable item identity; they count once. All new (sealed) supplement items were answered; the only failed requests (3 public items, GPU out-of-memory) are covered by the original P300 answers. Headline, rank, Composite and pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 69 | 46 | 43 | 36 | 26 | 41 | 21 | 22 | 44 | 27 | 22 | 14 | 24 | 28 | 11 | 27 | 27 | 25 | 43 | 33 | 38 | 13 | 2 |
| Respan Span-02self-hosted · S+P (draw v1.6-regular-20261012) | — | 60 | 50 | 57† | 61 | 80 | — | 53† | — | 37† | — | — | — | — | — | — | — | 70† | 29† | — | — | — | 72† |
| Seb-9B-v2self-hosted · not measured | — | 44 | 65 | 74† | 31 | 68 | — | 86† | — | 33† | — | — | — | — | — | — | — | 19† | 5† | — | — | — | 22† |
| wity-1 (Wity, reasoning off)API · S+P+L1+L2+L3 | 67 | 52 | 38 | 48 | 29 | 36 | 21 | 26 | 10 | 13 | 17 | 14 | 17 | 17 | 6 | 11 | 24 | 30 | 31 | 33 | 33 | 26 | 1 |
| APUS-OpenJev-v1-9Bself-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 74 | 48 | 33 | 31 | 25 | 37 | 18 | 22 | 37 | 41 | 8 | 0 | 20 | 25 | 6 | 16 | 28 | 29 | 30 | 34 | 35 | 17 | 12 |
| Arcbright Elf 1 12Bself-hosted · S+P (draw v1.6-regular-20261012) | — | 41 | 53 | 60† | 57 | 79 | — | 65† | — | 30† | — | — | — | — | — | — | — | 42† | 0† | — | — | — | 33† |
| Jeffan-31b-1710self-hosted · not measured | — | 76 | 71 | 58† | 67 | 77 | — | 69† | — | 45† | — | — | — | — | — | — | — | 76† | 58† | — | — | — | 67† |
| Plumb-4Bself-hosted · S+P+L1+L2+L3 | 72 | 45 | 35 | 45 | 36 | 30 | 38 | 27 | 32 | 34 | 25 | 8 | 32 | 40 | 14 | 24 | 31 | 25 | 28 | 27 | 39 | 22 | 20 |
| SPX-CD ProAPI · S+P+L1+L2+L3 | 82 | 70 | 44 | 55 | 37 | 44 | 53 | 43 | 62 | 50 | 35 | 47 | 41 | 19 | 22 | 47 | 37 | 48 | 54 | 29 | 52 | 29 | 47 |
| Messier One v0.2self-hosted · not measured | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |
| spark-s1-4b-v6self-hosted · S+P+L1+L2+L3 | 80 | 46 | 33 | 31 | 18 | 28 | 9 | 17 | 37 | 35 | 22 | 0 | 22 | 23 | 0 | 15 | 29 | 20 | 23 | 10 | 26 | 2 | 7 |
| KnowLine 31B-Gen1self-hosted · not measured | — | 81 | 72 | 68† | 75 | 88 | — | 63† | — | 46† | — | — | — | — | — | — | — | 80† | 53† | — | — | — | 67† |
| swanOneself-hosted · S+PThe pinned model weights cannot currently be accessed (Hugging Face returned HTTP 401). Supplement measurements remain pending access recovery; the radar shows the existing S + P items. | — | 57 | 86 | 68† | 48 | 85† | 26† | 45† | 91† | 42† | — | — | — | — | — | — | — | — | — | — | — | — | — |
| AJev lm6qself-hosted · not measured | — | 74 | 68 | 76† | 67 | 84 | — | 64† | — | 43† | — | — | — | — | — | — | — | 77† | 59† | — | — | — | 65† |
| Quyet-1.0-Mediumself-hosted · S+P+L1+L2+L3 | 60 | 44 | 33 | 32 | 25 | 42 | 25 | 15 | 35 | 32 | 23 | 8 | 15 | 20 | 0 | 29 | 33 | 35 | 19 | 21 | 35 | 12 | 14 |
| Bespoke Nimble-v4API · S+P (draw v1.6-regular-20261012) | — | 42 | 67 | 51† | 39 | 73 | — | 72† | — | 15† | — | — | — | — | — | — | — | 22† | 0† | — | — | — | 33† |
| Turion Jev 31Bself-hosted · not measured | — | 75 | 72 | 50† | 62 | 83 | — | 68† | — | 54† | — | — | — | — | — | — | — | 71† | 61† | — | — | — | 72† |
| Vansa-3.4API · A4+P+L1+L2+L3 | 54 | 47 | 12 | 42 | 28 | 17 | 12 | 10 | 24 | 43 | 11 | 0 | 13 | 19 | 12 | 9 | 17 | 26 | 15 | 33 | 35 | 0 | 11 |
| JPT-4Bself-hosted · S+P (draw v1.6-regular-20261012) | — | 36 | 75 | 63† | 27 | 75 | — | 70† | — | 30† | — | — | — | — | — | — | — | 11† | 14† | — | — | — | 33† |
| levself-hosted · S+P+L1+L2+L3 | 55 | 41 | 28 | 20 | 22 | 22 | 12 | 22 | 28 | 22 | 29 | 8 | 22 | 11 | 2 | 14 | 17 | 28 | 19 | 24 | 37 | 5 | 37 |
| NInfer Qwen3.8-Flash-Next mixedself-hosted · S+P+L1+L2+L3 | 72 | 53 | 41 | 50 | 24 | 33 | 30 | 33 | 55 | 33 | 28 | 16 | 43 | 20 | 27 | 36 | 35 | 54 | 31 | 38 | 43 | 25 | 20 |
| decider-4b v2self-hosted · S+P+L1+L2+L3 | 62 | 42 | 41 | 25 | 11 | 28 | 8 | 22 | 21 | 32 | 32 | 0 | 18 | 13 | 1 | 8 | 18 | 32 | 22 | 24 | 29 | 1 | 1 |
| baselabs mammoth-1self-hosted · not measured | — | 77 | 76 | 83† | 65 | 90 | — | 65† | — | 62† | — | — | — | — | — | — | — | 79† | 76† | — | — | — | 61† |
| Jeffan-4B-2010self-hosted · not measured | — | 39 | 66 | 76† | 30 | 63 | — | 83† | — | 14† | — | — | — | — | — | — | — | 20† | 0† | — | — | — | 22† |
| jev-localself-hosted · S+P+L1+L2+L3 | 71 | 46 | 43 | 31 | 21 | 31 | 21 | 22 | 33 | 32 | 17 | 7 | 20 | 12 | 7 | 21 | 29 | 10 | 34 | 25 | 29 | 11 | 5 |
| GPT-6 Luna (low reasoning effort)API · A4+P+L1+L2+L3L3 items were reviewed by an OpenAI model, so the L3-based language and category cells of this row carry that exposure; headline scores do not use L3. | 97 | 96 | 91 | 93 | 91 | 94 | 100 | 94 | 100 | 99 | 98 | 96 | 94 | 93 | 96 | 94 | 96 | 97 | 100 | 98 | 96 | 99 | 97 |
| Autoloops – Gemma 4 31B ITAPI · A4+P+L1+L2+L3 | 73 | 67 | 44 | 50 | 44 | 56 | 61 | 42 | 61 | 60 | 43 | 48 | 48 | 48 | 47 | 60 | 50 | 55 | 49 | 45 | 53 | 41 | 51 |
| Wald 4Bself-hosted · S+P (draw v1.6-fastlane-20261009) | 95† | 43 | 51 | 52† | 35 | 54 | 51† | — | 47† | 0† | — | — | — | — | — | — | 8† | 26† | 14† | — | — | — | 15† |
| GPT-6 Luna (default medium reasoning effort)API · A4+P+L1+L2+L3L3 items were reviewed by an OpenAI model, so the L3-based language and category cells of this row carry that exposure; headline scores do not use L3. | 97 | 97 | 98 | 96 | 91 | 98 | 100 | 99 | 99 | 96 | 98 | 100 | 96 | 95 | 98 | 95 | 97 | 99 | 95 | 96 | 100 | 99 | 96 |
| ClassOne Qwen 3.5 9Bself-hosted · S+P+L1+L2+L3Language and category cells use the unchanged original S+P observations and the newly completed frozen supplemental items (L1, L2, L3), answered with the original model and serving setup on an isolated pod. Each supplement carries the same original P300 answers by stable item identity; they count once in the union. Failed requests stay scored as failures and are not imputed. Headline, rank, Composite and original pricing are unchanged. L3 items were reviewed by an OpenAI model, so cells based on L3 carry that exposure; headline scores do not use L3. | 71 | 45 | 38 | 51 | 26 | 36 | 23 | 19 | 30 | 43 | 23 | 0 | 17 | 8 | 1 | 4 | 11 | 30 | 17 | 18 | 40 | 13 | 31 |
| JevK5 v0.3self-hosted · S+P+L1+L2+L3 | 67 | 43 | 40 | 42 | 29 | 19 | 20 | 13 | 28 | 30 | 18 | 0 | 28 | 27 | 8 | 16 | 28 | 17 | 22 | 34 | 29 | 3 | 18 |
| metask-jev-4bself-hosted · S+P+L1+L2+L3 | 73 | 40 | 30 | 34 | 15 | 21 | 16 | 19 | 28 | 24 | 12 | 9 | 15 | 4 | 0 | 9 | 11 | 23 | 16 | 12 | 26 | 4 | 0 |
| deck-4B v1.0self-hosted · S+P+L1+L2+L3 | 67 | 43 | 37 | 45 | 32 | 22 | 21 | 19 | 29 | 30 | 22 | 0 | 31 | 26 | 5 | 13 | 23 | 18 | 27 | 36 | 29 | 19 | 17 |
| GLiClass v4 9Bself-hosted · not measured | — | 36 | 58 | 65† | 59 | 80 | — | 71† | — | 45† | — | — | — | — | — | — | — | 14† | 0† | — | — | — | 39† |
| Kahn1 4Bself-hosted · S+P+L1+L2+L3 | 67 | 40 | 36 | 46 | 28 | 35 | 25 | 30 | 29 | 45 | 25 | 5 | 35 | 25 | 21 | 14 | 33 | 19 | 28 | 41 | 30 | 20 | 20 |
| Clef-Flashself-hosted · S+P+L1+L2+L3 | 69 | 44 | 40 | 31 | 8 | 34 | 9 | 19 | 41 | 29 | 23 | 18 | 10 | 2 | 0 | 11 | 12 | 23 | 23 | 23 | 36 | 13 | 11 |
| Decision-4Bself-hosted · S+P+L1+L2+L3 | 61 | 43 | 27 | 19 | 0 | 31 | 0 | 7 | 20 | 22 | 26 | 2 | 6 | 10 | 0 | 0 | 21 | 17 | 13 | 15 | 25 | 0 | 0 |
| decisor-4bself-hosted · S+P (draw v1.6-fastlane-20261009) | 100† | 36 | 70 |