AyaVisionBench
Aya Vision 32B 64% Dropped Vision-language models answer image questions in 23 languages across 9 task categories, 135 image-question pairs per language. A model judge compares two models' answers and reports win rates.
Images come from the Cauldron held-out test set. Questions were human-checked in English and machine-translated into 22 other languages.
- Level
- Everyday Most adults, without training
- Scored
- Graded by a model A language model scores each response, usually against a rubric.
- Answers
- Open-ended Free-form text, documents, designs or media judged as a whole.
Aya Vision 32B leads with 64%
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Aya Vision 32B | 64% | Mar 2025 | Model card |
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
No launch post has listed it since Mar 2025
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| 4 Mar 2025 | Aya Vision 32B | 64 | Up to 64% win rate on AyaVisionBench (best pairwise win rate against Llama-3.2 90B Vision, Molmo 72B, Qwen2-VL 72B) |
One system card reports it
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| 13 May 2025 | Aya Vision 32B technical report | AyaVisionBench, pairwise win rates averaged over 23 languages, no overall: vs Llama-3.2-90B-Vision 65.9, vs Molmo-72B 61.2, vs Qwen-2.5-VL-72B 48.5 (Figure 7, printed labels) | Capabilities |
More in Vision
- MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy. GPT-5.1 85.4%
- MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. Kimi K3.1 Pro 84.1%
- CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
- MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy.
Qwen 3.8 Max 97.7% - ChartQA In use Models answer human-written and generated questions about chart images that need visual and arithmetic reasoning. Scored by relaxed accuracy with a 5 percent numeric tolerance. Command A Vision 90.9%
- DocVQA In use Models answer 50,000 questions about 12,000 scanned document images by reading their text and layout. Scoring is ANLS, a string similarity to reference answers.
Seed1.5-VL 96.9%