decision-index-leaderboard

# Decision Index The text board of the [Jev Decision Index](https://huggingface.co/spaces/multimodalart/jev-decision-index) (Decision Index 0.4.1): models that read a state and a typed question and return a decision. Vision board: [multimodalart/decision-index-vision-leaderboard](https://huggingface.co/datasets/multimodalart/decision-index-vision-leaderboard). Tasks (each is its own leaderboard): - `decision_index`: the Decision Index (Full) = 0.6 × same-skills + 0.4 × new-domains, both equated on 21 stock base models. Scores within 0.9 points share a rank. - `same_skills`: private sets that test the same skills as the public suite. - `new_domains`: private sets from domains outside the public suite. - `public`: the 42-benchmark public suite. The private and new-domain sets are not released; only aggregate scores are published. The evaluation kit and submission process: [apolinario/decision-index](https://github.com/apolinario/decision-index). Methodology: https://huggingface.co/spaces/multimodalart/jev-decision-index. This repo mirrors the board: `leaderboard` has one row per model (248 entries), `benchmarks` the per-benchmark public scores. It is regenerated whenever the board is updated. ## Top 10 | Rank | Model | Full | |---|---|---| | 1 | KnowLine 31B Gen1 | 71.86 | | 2 | Fastino GLiDE 1.2 | 71.52 | | 3 | Sieve 31B v3 | 71.10 | | 4 | Turion Jev 31B | 70.47 | | 5 | vLLM-SR d3-ultra (Decision 3.0, 27B) | 70.18 | | 6 | Jeffan-31b-1710 | 69.75 | | 7 | H2O Lightning 31B v1.1 | 68.64 | | 8 | WeMM-JEV-27B | 68.49 | | 9 | baselabs mammoth-1 (27B) | 68.44 | | 10 | Ainvest-Decision-27B v1.0 | 68.23 | Generated 2026-10-29T14:04:16Z.