JevBench coherence Leaderboard
decision-diagnostic21 models · Updated October 2026
On the JevBench coherence benchmark, Open-Jev-9B ranks #1 with a score of 83.9. The full ranking, with cost per million tokens, is below.
| # | Model | Score | ECE | p50 latency | Input / 1M |
|---|---|---|---|---|---|
| 1 | Open-Jev-9BZefanCai | 83.9 | — | — | — |
| 2 | JevAny-Gemma-4B-LoRASimpleJev | 81.2 | — | — | — |
| 3 | JevAny-Qwen3.5-4B-LoRASimpleJev | 79.4 | — | — | — |
| 4 | Decider 4BMapika | 78.4 | — | — | — |
| 5 | JevAny-Qwen3.5-4B-Direct-Token-LoRASimpleJev | 77.8 | — | — | — |
| 6 | TinyJev-4BAnkitAI | 77.3 | — | — | — |
| 7 | jebadiah-9b-v2Frontier Infra | 76.9 | — | — | — |
| 8 | Vonwfzyx | 75.9 | — | — | — |
| 9 | JevK5 v0.3alibiserikbay | 72.8 | — | — | — |
| 10 | OneJev-4BOmniJev | 72.5 | — | — | — |
| 11 | rsi-jev-v3.0-qwen3.5-2bshgao | 72.2 | — | — | — |
| 12 | Kev 4B r10Jared Palmer | 71.9 | — | — | $0.04 |
| 13 | Decider 2BMapika | 71.7 | — | — | — |
| 14 | jevhome-Ejevhome | 69.6 | — | — | — |
| 15 | Bespoke-Nimble-9BBespoke Labs | 69.2 | — | — | — |
| 16 | CLM-v0.1-8BContrastive-LM | 66.8 | — | — | — |
| 17 | OpenThai-SystemOneiapp | 64.2 | — | — | — |
| 18 | jevhome-ettin-1bjevhome | 64.1 | — | — | — |
| 19 | LAYA typed-decisionsConvai Innovations | 63.9 | — | — | — |
| 20 | RLCD ModernBERT 151Mheman10x | 59.4 | — | — | — |
| 21 | LAYAConvai Innovations | 56.5 | — | — | $0.00 |
Frequently asked questions
What is the JevBench coherence benchmark?
JevBench coherence is a decision-diagnostic used to evaluate and rank large language models. This page lists 21 models by their JevBench coherence score.
Which model is best on JevBench coherence?
Open-Jev-9B (ZefanCai) currently leads the JevBench coherence leaderboard with a score of 83.9.
What is the best value model on JevBench coherence?
Pricing data is being matched to these models.
Pricing is indicative — confirm with the provider before production use. Updated October 2026.