FireWorldBench - Causal Mechanism Attribution (Rendered Images): leaderboard

Metric: Completion accuracy (%; P4, causal mechanism attribution 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 2030. 10 models tracked.

Top models

#ModelScore
1Gemini 2.5 Flash41.12
2Qwen 3 VL 32B Instruct36.91
3InternVL3-38B35.55
4GPT-5.6 Luna34.4
5Qwen 3 VL 235B A22B Instruct34.38
6Claude Haiku 4.534.03
7InternVL3-78B33.42
8Qwen 3.5 Plus (2026-02-15)32.75
9InternVL3-8B26.59
10Qwen 3 VL 8B Instruct24.8

Interactive version: theaggregate.ai/benchmark?slug=fireworldbench-causal-mechanism-attribution-rendered-images · How It Works · Data refreshed daily, snapshot 2026-09-26.