FireWorldBench - Counterfactual Intervention Reasoning (Rendered Images): leaderboard

Metric: Completion accuracy (%; P5, counterfactual intervention reasoning 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 Flash43.46
2InternVL3-78B41.22
3Qwen 3.5 Plus (2026-02-15)37.83
4Qwen 3 VL 8B Instruct36.42
5Qwen 3 VL 32B Instruct35.82
6GPT-5.6 Luna34.61
7Claude Haiku 4.532.47
8InternVL3-38B31.94
9InternVL3-8B26.27
10Qwen 3 VL 235B A22B Instruct20.62

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