DiningBench - VQA: leaderboard
DiningBench (2026) evaluates vision-language models on multi-view user photos of restaurant dishes, testing dish classification against same-menu distractors, nutrition estimation and food visual question answering. This board is accuracy in percent on food visual questions about cooking technique, dietary advice and counterfactual or multi-image reasoning, judged against references; higher is better.
Metric: Accuracy (%) on the 804 DiningBench visual questions (cuisine technique, dietary advice, counterfactual and multi-image reasoning), answers judged by an LLM against the reference; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 29 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 90.42 | #119 |
| 2 | Gemini 2.5 Pro | 89.93 | #276 |
| 3 | Gemini 3 Flash (Preview) | 88.56 | #152 |
| 4 | GPT-5 | 86.94 | #178 |
| 5 | GPT-4.1 | 84.33 | #505 |
| 6 | Claude Sonnet 4.5 | 83.58 | #261 |
| 7 | GPT-4o | 80.6 | #751 |
| 8 | Qwen 3 VL 30B A3B Instruct | 80.6 | #895 |
| 9 | O4 Mini | 80.35 | #368 |
| 10 | Gemini 2.5 Flash | 78.36 | #463 |
| 11 | Qwen 2.5 VL 72B Instruct | 76.62 | #841 |
| 12 | MiMo-VL-7B-RL | 75 | #1155 |
| 13 | Qwen 3 VL 30B A3B (Thinking) | 74.13 | #719 |
| 14 | Qwen 3 VL 8B Instruct | 72.76 | #1029 |
| 15 | InternVL3.5-38B | 72.51 | #1021 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=diningbench-vqa · How It Works · Data refreshed daily, snapshot 2026-11-21.