MechVQA - Consistency Judgment: leaderboard
Metric: Accuracy (%) on the Consistency Judgment subtask (Judging capability), MechVQA test split (drawing-level 8:1:1 split of 20,778 questions on 3,281 mechanical drawings), answers judged against the reference by three LLM judges (GPT-OSS-120B, DeepSeek-V3.2, Kimi-k2); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 82.67 |
| 2 | GPT-5 | 70.67 |
| 3 | Qwen 3 VL 32B Instruct | 69.33 |
| 4 | GLM-4.6V | 69.33 |
| 5 | GPT-4o | 66.67 |
| 6 | Claude Sonnet 4.5 | 64 |
| 7 | Qwen 3 VL 30B A3B Instruct | 57.33 |
| 8 | GPT-4o Mini | 49.33 |
| 9 | Qwen 3 VL 4B Instruct | 49.33 |
| 10 | Gemma 3 27B (IT) | 40 |
| 11 | Llama 3.2 11B Instruct | 28 |
Interactive version: theaggregate.ai/benchmark?slug=mechvqa-consistency-judgment · How It Works · Data refreshed daily, snapshot 2026-10-07.