FireWorldBench - Cross-Field Coupling Understanding (Structured Records): leaderboard
The cross-field coupling understanding dimension of FireWorldBench (Chen et al., 2026): relating coupled fields of a simulated fire, such as plume shape and flow, to predict smoke transport. Observations are given as structured sensor traces and physical records; the score is completion accuracy in percent, averaged over matched multiple-choice and open-report questions and scored deterministically; higher is better.
Metric: Completion accuracy (%; P3, cross-field coupling understanding 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 April 2028. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Luna | 55.49 | #117 |
| 2 | Gemma 3 27B (IT) | 46.56 | #1306 |
| 3 | DeepSeek V3.2 | 45.12 | #403 |
| 4 | Kimi K2 | 44.53 | #435 |
| 5 | Qwen 3 235B A22B | 43.73 | #622 |
| 6 | Qwen 3.5 Plus (2026-02-15) | 42.76 | #178 |
| 7 | Gemini 2.5 Flash | 42.52 | #452 |
| 8 | Qwen 3 32B | 40.16 | #960 |
| 9 | Claude Haiku 4.5 | 39.92 | #568 |
| 10 | GLM 4.5 Air | 39.84 | #863 |
| 11 | Qwen 3 Next 80B A3B Instruct | 39.53 | #763 |
| 12 | Llama 3.3 70B Instruct | 37.11 | #1353 |
| 13 | Qwen 3 8B | 33.88 | #1838 |
| 14 | Gemma 3 12B (IT) | 33.3 | #1849 |
| 15 | Llama 3.1 8B Instruct | 28.38 | #4277 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=fireworldbench-cross-field-coupling-understanding-structured-records · How It Works · Data refreshed daily, snapshot 2026-11-12.