Crucible
Crucible
decisions per simulated day
blocked or invalid actions
| Benchmark | Score | Outcomes | Crises | Discipline | Fidelity | Decisions | Rules | Effort |
|---|---|---|---|---|---|---|---|---|
| Crucible | 71 (floor 26) | 71 | 100 | 40 | 30 | 17 | +39 | api-default |
Floors show what a do-nothing baseline earns on that scenario — read every score against its floor. Runs without a comparison field are solo measurements.
deepseek-v4-flash-0731 · fable · fixture · gemma-4-26b-a4b-it-qat-mlx · google/gemma-4-31b-it · gpt-5.6-sol · k3 · kimi-code/k3 · meta/llama-3.2-3b-instruct · meta/llama-3.3-70b-instruct · meta/llama-4-maverick-17b-128e-instruct · minimax/minimax-m2.5 · mistralai/mixtral-8x22b-v0.1 · nvidia-nemotron-3.5-lightning-30b-a3b · openai/gpt-oss-120b · opus · ornith-1.5-35b-a3b-mlx@6bit · ornith-1.5-9b-mlx · qwen3.8-27b · sonnet