FireWorldBench - Physical Field Perception and Grounding (Rendered Images): leaderboard
The physical field perception and grounding dimension of FireWorldBench (Chen et al., 2026): grounding partial observations of a simulated fire in the underlying physical fields, such as locating a source from temperature and soot anomalies. Observations are given as physics-rendered field images and 3D scene renderings; 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 (%; P2, physical field perception and grounding items of the controlled-simulation subset; image-observation track (physics-rendered field visualizations and 3D scene renderings); 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 May 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Luna | 41.11 | #117 |
| 2 | Gemini 2.5 Flash | 40.84 | #452 |
| 3 | Qwen 3.5 Plus (2026-02-15) | 40.77 | #178 |
| 4 | Qwen 3 VL 235B A22B Instruct | 40.42 | #521 |
| 5 | Qwen 3 VL 32B Instruct | 40.23 | #582 |
| 6 | InternVL3-78B | 37.17 | #733 |
| 7 | Claude Haiku 4.5 | 34.52 | #568 |
| 8 | InternVL3-38B | 26.66 | #905 |
| 9 | InternVL3-8B | 25.08 | #1695 |
| 10 | Qwen 3 VL 8B Instruct | 21.47 | #982 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=fireworldbench-physical-field-perception-and-grounding-rendered-images · How It Works · Data refreshed daily, snapshot 2026-11-12.