FoodMonitor - Environment Violations: leaderboard

FoodMonitor Environment Violations scores how well multimodal models detect scene-level food-safety violations, such as dirty work surfaces, uncovered waste bins or standing water, in 477 kitchen videos with a 27-item rule document (FoodMonitor, 2026). The score is detection F1 on a 0 to 100 scale, higher is better.

Metric: Environment-violation F1 (0-100; the paper's 0-1 score times 100), on the 477 one-minute FoodMonitor kitchen surveillance clips (3,307 annotated violations of a 27-item food safety rule document), frames at 1 fps plus the full rule document, structured JSON output; a prediction counts only if its rule category matches and a text comparison model accepts its description as the same violation, one-to-one; all models in thinking mode; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 11 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Seed 2.0 ProThinking45.9#222
2Doubao Seed 2.0 LiteThinking42.1#204
3GLM-4.6V38.2#710
4Gemini 3 Flash37.6#175
5Qwen 3.5 397B A17B32.1#264
6Qwen 3 VL 235B A22B (Thinking)30.3#459
7Gemini 3 Pro29.7#145
8Kimi K2.525.4#266
9GLM-4.6V-Flash24.2#1040
10Qwen 3 VL 8B (Thinking)20.4#910
11MiMo-VL-7B-RL18#1155

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=foodmonitor-environment-violations · How It Works · Data refreshed daily, snapshot 2026-11-21.