FireWorldBench - Causal Mechanism Attribution (Structured Records): leaderboard
Metric: Completion accuracy (%; P4, causal mechanism attribution items of the controlled-simulation subset; structured-observation track (textual sensor traces and physical records); each cell is the mean of the matched choice-format score (Jaccard overlap of selected and gold option sets) and open-report score (exact agreement of required structured slots); deterministic evaluator, no LLM judge; failed or unparseable responses score 0). Source: arxiv.org. Saturation forecast: Around 2031. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Luna | 52.14 |
| 2 | DeepSeek V3.2 | 46.91 |
| 3 | Qwen 3 8B | 42.98 |
| 4 | Qwen 3 32B | 42.67 |
| 5 | Claude Haiku 4.5 | 42.03 |
| 6 | Kimi K2 | 40.98 |
| 7 | Qwen 3.5 Plus (2026-02-15) | 40.61 |
| 8 | Qwen 3 235B A22B | 40.56 |
| 9 | Llama 3.3 70B Instruct | 38.02 |
| 10 | GLM 4.5 Air | 37.49 |
| 11 | Gemini 2.5 Flash | 37.32 |
| 12 | Llama 3.1 8B Instruct | 37.16 |
| 13 | Gemma 3 27B (IT) | 37.08 |
| 14 | Gemma 3 12B (IT) | 34.03 |
| 15 | Qwen 3 Next 80B A3B Instruct | 32.91 |
Interactive version: theaggregate.ai/benchmark?slug=fireworldbench-causal-mechanism-attribution-structured-records · How It Works · Data refreshed daily, snapshot 2026-09-26.