MechVQA - Text and Table: leaderboard

MechVQA (2026) tests multimodal LLMs on understanding mechanical engineering drawings, with 20,778 questions on 3,281 drawings scored as accuracy by three LLM judges against reference answers. This board covers the Text and Table subtask of the recognition capability, as accuracy in percent on the test split; higher is better.

Metric: Accuracy (%) on the Text and Table subtask (Recognition 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: Estimated already saturated. 15 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Claude Sonnet 4.597.73#267
2Gemini 3 Pro (Preview)97.73#123
3Qwen 3 VL 32B Instruct97.73#607
4GLM-4.6V97.73#721
5Qwen 3 VL Plus95.45#514
6GPT-593.18#185
7MiMo-VL-7B-SFT93.18#1097
8Qwen 3 VL 30B A3B Instruct90.91#903
9Qwen 3 VL 4B Instruct88.64#1348
10MiniCPM-V-4.584.09#1451
11GPT-4o81.82#754
12InternVL3.5-8B77.27#1618
13Gemma 3 27B (IT)70.45#1341
14GPT-4o Mini68.18#1680
15Llama 3.2 11B Instruct54.55#5213

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=mechvqa-text-and-table · How It Works · Data refreshed daily, snapshot 2026-11-22.