Hosting = where inference runs; company = where the provider or lab is registered.
EU hosting means the route's inference runs inside the EU: an EU region, AWS Bedrock's EU cross-region (geo) profiles, Azure's Europe Data Zone, or a provider whose entire public fleet is documented as EU-hosted, each checked per model against the provider's documentation. Global deployments do not count, and neither does an EU billing region, an EU company or an EU control plane on its own. One disclosed company-policy exception stays in, marked “EU equivalent”.
Evidence requirements (benchmark evidence, priced provider, measured task tokens) sit above the table they apply to.
Applies to price views & model offers; benchmark evidence stays unfiltered.
Data: JevBench v1.6.1, published 2026-10-06
Capability rank #6 in v1.6.1.
Compare Jev-compatible models by separate measures and their measurement conditions. A historical measurement is not a current rank.
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.
Low sample, n < 30 — indicative only
| Category (items) | A: Jev 1.13.0 | B: Jev-Omni |
|---|---|---|
| Feature extraction for predictive modeling (19) | 22.5 n=19 | 41.4 n=19 |
| Graphs and knowledge graphs (18) | 49.8 n=18 | 42.2 n=18 |
| LLM guardrails (17) | 44.1 n=17 | 65.6 n=17 |
| Search and retrieval (17) | 80.5 n=17 | 80.9 n=17 |
| Lead generation (17) | 67.5 n=17 | 80.0 n=17 |
| Semantic code linting (16) | 12.2 n=16 | 9.8 n=16 |
| Demand forecasting (16) | 0.0 n=16 | 0.0 n=16 |
No published value for either system (under 15 answered items, or no per-category values): Gaming (14), Scientific discovery (14), Recruiting (13), Advertising (11).
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.
Not drawn: Other — none of the above: general reasoning, arithmetic or date puzzles, everyday assistant requests without a business workflow. 145 items (24 open / 121 sealed) — not a use case of its own, so it is counted but not drawn.
Chance-corrected competence (0 = chance, 100 = perfect) over the category's items the system saw, open and sealed pooled, tiers pooled: per request type the method's competence cell, then the mean over the request types present. Not part of the JevBench Score; never equated for API systems; compare systems within a category, not categories with each other. Values can be negative (below chance).
Family and language are authoring metadata of every item in the frozen v1.6 pool. Each of the 1,500 v1.6 items was also labelled with one subject topic (the v1.2 topic list, unchanged) and one TypeSafe use-case category by Winnow-12B Q8 on our own GPU pod (same model, questions and taxonomy as v1.5); where a uc1 item was written for a use case, that authoring use case is used instead. Sealed items stayed on our own infrastructure. 75 public items (5 %) were checked by hand; item-group rules fix the systematic misses (topic agreement before rules 93 %). All non-English uc1 items are machine-authored and not native-reviewed. Self-hosted systems saw S u P (1,500 items); hosted API systems saw only A u P (600 items), so their cells cover fewer items.
| Spoke | A: Jev 1.13.0 | B: Jev-Omni |
|---|---|---|
| The four score axes | ||
| Intelligence | 63.6 | 55.5 |
| Calibration | 90.6 | 87.0 |
| Speed | 91.5 | 85.4 |
| Cost | 54.7 | 56.1 |
| Capability by subject topic | ||
| Rules, policy & law | 70.6 | 58.7 |
| Coding & software | 74.5 | 63.0 |
| Math & numbers | 30.5 | 23.3 |
| Support & operations | 34.9 | 39.9 |
| Finance & commerce | 60.3 | 76.6 |
| Everyday language | 51.5 | 32.4 |
| Safety & security | 35.2 | 63.3 |
| Use cases (TypeSafe categories) | ||
| Model routing | 78.2 | 69.1 |
| Legal & compliance | 81.3 | 60.5 |
| Customer support | 47.6 | 34.5 |
| E-commerce | 44.6 | 52.9 |
| Risk assessment | 42.1 | 43.2 |
| Insurance claims | 76.5 | 78.6 |
| Financial crime | 36.0 | 34.5 |
| Moderation | 53.4 | 80.3 |
| Competence per request type, open / sealed | ||
| Choice · open | 79.0 | 75.0 |
| Choice · sealed | 75.8 | 68.9 |
| Noul · open | 54.5 | 45.1 |
| Noul · sealed | 46.1 | 32.5 |
| Score · open | 63.3 | 57.3 |
| Score · sealed | 63.0 | 54.4 |
| Competence per tier — open set | ||
| Easy | 93.9 | 83.9 |
| Standard | 63.2 | 58.9 |
| Judge | 67.7 | 51.3 |
| Hard | 65.6 | 66.8 |
| Competence per tier — sealed set | ||
| Easy | 78.4 | 67.6 |
| Standard | 68.7 | 61.2 |
| Judge | 66.9 | 60.6 |
| Hard | 58.9 | 48.0 |
| Measure | Jev 1.13.0 (TypeSafe AI) | Jev-Omni (akhilaaa3, Gemma-4-12B merged) |
|---|---|---|
| Capability Score | 77.1 | 71.3 |
| Capability rank | #4 | #6 |
| Composite (secondary) | 71.5 | 67.7 |
| intelligence | 63.6 | 55.5 |
| calibration | 90.6 | 87.0 |
| speed | 91.5 | 85.4 |
| cost | 54.7 | 56.1 |
| USD per 1,000 decisions | $0.032 per 1,000 decisions | $0.029 per 1,000 decisions |
| p50 latency (seconds) | 0.2 | 0.5 |
| Last measured | 2026-10-05 | 2026-10-02 |
proprietary API
Published sourceApache-2.0, following Gemma 4; dataset rights stated separately by the author
Published source