whatmodel.sh · find the best decision models you can run

Best Decision models for probability consistency

A leaderboard of the best Decision models for probability consistency. JevAny-Gemma-4B (pointer) (JevBench mini) is #1 right now, followed by Open-Jev-9B (JevBench mini) and JevAny-Qwen-4B (pointer) (JevBench mini). JevBench mini probability-law pass rates. These measure coherence, not correctness; see the separate accuracy measurement on model pages. Filter by your computer or budget to find the best one you can use.

24 models · 5 benchmarks · updated

Compare tested decision models and inference recipes across text, vision and workflow tasks. Use task filters and raw results: probability consistency measures coherence, not correctness.

Filter by your hardware and budget · reset
price[local][any]max per M input tokens
Filter by task

Scores use reviewed consistency evaluations only. Compare scores within this task. Models without task measurements are excluded.

New #1s:
orRSS

detecting your OS…

$ terminal
#modeltask score↓where↓RAM↓speed↓get itsources
1JevAny-Gemma-4B (pointer) (JevBench mini) SimpleJev67±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
2Open-Jev-9B (JevBench mini) Zefan Cai67±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
3JevAny-Qwen-4B (pointer) (JevBench mini) SimpleJev67±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
4decider-4b v2.1 (JevBench mini) Mapika66±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
5JevAny-Qwen-4B (direct token) (JevBench mini) SimpleJev65±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
6Bespoke-Nimble-9B (JevBench mini) Bespoke Labs63±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
7Jebadiah 9B v2 (JevBench mini) Frontier Infra63±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
8OneJev-4B (JevBench mini) OmniJev58±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
9TinyJev-4B (JevBench mini) AnkitAI57±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
10JevK5 v0.3 (JevBench mini) Alibi Serikbay57±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
11decider-2b v10 (JevBench mini) Mapika53±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
12OpenThai-SystemOne (JevBench mini) iApp Technology53±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
13RSI-Jev v3.0 2B (JevBench mini) Shanghua Gao46±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
14Kev-4B (JevBench mini) Jared Palmer45±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
15jevhome-ettin-1b (JevBench mini) Jibril Frej41±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
16Von 1.3 (JevBench mini) wfzyx39±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
17jevhome-E (JevBench mini) Jibril Frej39±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
18CLM v0.1 (JevBench mini) Contrastive-LM32±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
19Verdict 1.4 (JevBench mini) Heman10x31±5source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
20Laya (typed-decisions) (JevBench mini) Convai Innovations21±6source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
21Laya (english) (JevBench mini) Convai Innovations19±6source code ↗github.com ↗JevBench 0.1.0 mini evaluation reports
—decider-2b (DecisionBench fa996cea) Mapikanot yet calibratedmodel files ↗huggingface.co ↗Chao Ma · independent injection study
—Jev 1.13 TypeSafenot yet calibratedapi $0.042/M input tokensdocs.typesafe.ai ↗Chao Ma · independent injection study
—open-jev-deberta-v3-large (DecisionBench 19bf9a64) com-kotobalabsnot yet calibratedmodel files ↗huggingface.co ↗Chao Ma · independent injection study

24 of 24 catalogue models listed here · sources and licences

* new or thinly tested: ranked on too few benchmark results · † no independently run results verified · ~RAM is estimated; full context may need more memory