Gemini 3.1 Flash-Litev1.7 · 10 s rule
Data: JevBench v1.8.7, published 2026-11-03 · method v1.7 (10 s rule), re-scored from the recorded per-decision times
Capability 59.8 · API offering, ranked on the API leaderboard · outside the Jev-class caps (cost 7.1× Jev (v1.5 reference)), below the Capability ranking
Measured 2026-10-06 on A4 ∪ P, 600 items, equated (+2.57 I / +3.35 C); round score-a4-2.
proprietary API
intelligence
58.6
calibration
60.9
speed
79.0
cost
29.2
Composite (secondary)
16.9 · Composite rank #39 of 48 on the API leaderboard
Cost and measurement conditions
$0.230 per 1,000 decisions (public tariff)
measured: token usage of the 6 Oct 2026 A4 u P run x tariff USD 0.25/M input, 1.5/M output (spec gemini-3.1-flash-lite.json), per 1,000 decisions on the v1.6.1 common cost basis (591 of 600 items; the 9 A4 u P items outside the Jev reference's accepted input range are excluded, cost-basis-a4.json)
p50 latency: 1.0 s
none (API)
Measured latency over 600 decisions: p50 0.962 s · p95 1.317 s
0 decisions over 10 s.
A decision that takes longer than 10 seconds (measured time) counts as a wrong answer and as zero calibration for that item: each type's calibration is multiplied by the share of its decisions answered within 10 s. The limit uses the measured time (raw). The displayed adjusted time for self-hosted rows (2 × + 0.15 s) is an assumption and is not used for the limit.
JevBench v1.6.43 values (method v1.6, before the 10 s rule)
Capability 59.75 · Composite 16.94 · rank #148 on the all-systems board · #37 API
intelligence 58.57 · calibration 60.93 · speed 79.03 · cost 29.17
v1.6 speed basis (124 items): p50 0.957 s · p95 1.306 s
These figures are kept for comparison. The current figures above are re-scored under the 10 s rule from the recorded per-decision times and replace them.
Breakdowns (not part of the score)
Chance-corrected competence (0 = chance, 100 = perfect) on A4 ∪ P (600 items: 300 open + 300 sealed; fewer sealed items than the full set, so wider uncertainty). Raw values, not equated. Category cells use A4+P+L1+L2+L3; per-type/tier cells use the original basis above. Radars against Jev.
| Request type | Open | Sealed |
|---|---|---|
| choice | 70.9 | 66.1 |
| noul | 63.0 | 56.3 |
| score | 57.7 | 21.9 |
Capability by subject topic
| Rules, policy & law | 53.5 | n=1133 |
|---|---|---|
| Coding & software | 69.9 | n=202 |
| Math & numbers | 23.9 | n=269 |
| Finance & commerce | 59.3 | n=178 |
| Support & operations | 69.8 | n=109 |
| Everyday language | 84.8 | n=130 |
| Safety & security | 68.7 | n=84 |
Use cases (TypeSafe categories)
| Model routing | 73.8 | n=267 |
|---|---|---|
| Legal & compliance | 58.9 | n=269 |
| Customer support | 62.3 | n=156 |
| E-commerce | 41.3 | n=88 |
| Insurance claims | 39.1 | n=92 |
| Risk assessment | 63.9 | n=88 |
| Financial crime | 33.8 | n=76 |
| Feature extraction | 44.6 | n=90 |
| Lead generation | 66.7 | n=86 |
| Recruiting | 42.8 | n=88 |
| Knowledge graphs | 44.3 | n=82 |
| LLM guardrails | 61.6 | n=82 |
| Moderation | 75.0 | n=69 |
| Code linting | 61.2 | n=81 |
| Search & retrieval | 82.0 | n=84 |
| Science | 34.1 | n=82 |
| Advertising | 44.4 | n=85 |
| Gaming | 35.7 | n=79 |
| Demand forecasting | 4.4 | n=72 |
Cells need at least 15 answered items; † fewer than 30 (indicative only, listed below the radar in the compare view).
Previous measurement (v1.5.x, not comparable)
Capability 59.8 · composite 16.9 · cost $0.230 per 1,000 decisions (public list price)
Last measured 2026-10-06, measurement release v1.6.1. Older method and scale; the current figures above replace it.