Benchmark Almanac

xMMMU

GPT4o (snapshot unspecified) 58.3% tops its leaderboard Llama 3.2 11B 41.4% leads the models we track
Aya Vision 32B 45.1% leads since May 2025
Only in system cards

Vision language models answer college-level MMMU questions translated into 7 languages. Scored by accuracy.

Released as part of PangeaBench with the Pangea model.

Released
October 2024
Kind
General visual questions
Level
University
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Last in a system card
May 2025
Last model added
May 2025
Scores refreshed
November 12th, 2026
14
Entries on its own leaderboard, 5 models we track
1
System card reports it, from 1 lab

GPT4o (snapshot unspecified) tops the leaderboard with 58.3%

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Llama 3.2 11B, at 41.4%.

Aya Vision 32B leads with 45.1%, claims included

Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model.

The newest model with a score here came out in September 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

14 entries. Scroll sideways for the rest.

  1. 58.3 GPT4o (snapshot unspecified)
  2. 57.7 Gemini-1.5-Pro (snapshot unspecified)
  3. 43.7 Pangea-7B (Ours)
  4. 41.4 Llama 3.2 11B
  5. 41.0 LLaVA-OneVision 7B
  6. 40.4 Molmo-7B-D
  7. 38.8 Phi-3.5-vision-instruct
  8. 34.3 Llava-Next-7B
  9. 33.2 Cambrian-1-8B
  10. 31.5 Llava-1.5-7B
  11. 30.8 mBLIP BLOOMZ
  12. 30.5 PALO-7B
  13. 30.4 mBLIP mT0-XL
  14. 25.2 PaliGemma-3B
  1. 1 GPT4o (snapshot unspecified) 58.3%
  2. 2 Gemini-1.5-Pro (snapshot unspecified) 57.7%
  3. 3 Pangea-7B (Ours) 43.7%
  4. 4 Llama 3.2 11B 41.4%
  5. 5 LLaVA-OneVision 7B 41%
  6. 6 Molmo-7B-D 40.4%
  7. 7 Phi-3.5-vision-instruct 38.8%
  8. 8 Llava-Next-7B 34.3%
  9. 9 Cambrian-1-8B 33.2%
  10. 10 Llava-1.5-7B 31.5%
  11. 11 mBLIP BLOOMZ 30.8%
  12. 12 PALO-7B 30.5%
  13. 13 mBLIP mT0-XL 30.4%
  14. 14 PaliGemma-3B 25.2%

Source: Official xMMMU leaderboard

Sources: Labs’ model cards and official xMMMU leaderboard

#EntryScoreSettingDateSource
1 GPT4o (snapshot unspecified) 58.3% Official leaderboard
2 Gemini-1.5-Pro (snapshot unspecified) 57.7% Official leaderboard
3 Pangea-7B (Ours) 43.7% Official leaderboard
4 Llama 3.2 11B 41.4% Sep 2024 Official leaderboard
5 LLaVA-OneVision 7B 41% Aug 2024 Official leaderboard
6 Molmo-7B-D 40.4% Sep 2024 Official leaderboard
7 Phi-3.5-vision-instruct 38.8% Aug 2024 Official leaderboard
8 Llava-Next-7B 34.3% Official leaderboard
9 Cambrian-1-8B 33.2% Jun 2024 Official leaderboard
10 Llava-1.5-7B 31.5% Official leaderboard
11 mBLIP BLOOMZ 30.8% Official leaderboard
12 PALO-7B 30.5% Official leaderboard
13 mBLIP mT0-XL 30.4% Official leaderboard
14 PaliGemma-3B 25.2% Official leaderboard
#ModelScoreSettingDateSource
1 Aya Vision 32B 45.1% May 2025 Lab model card
2 Llama 3.2 11B 41.4% Sep 2024 Official leaderboard
3 LLaVA-OneVision 7B 41% Aug 2024 Official leaderboard
4 Molmo-7B-D 40.4% Sep 2024 Official leaderboard
5 Phi-3.5-vision-instruct 38.8% Aug 2024 Official leaderboard
6 Cambrian-1-8B 33.2% Jun 2024 Official leaderboard

Labs report scores for 1 model in their cards. Switch the scores above to rank them together.

Number one over time 2 models have held number one with claims included since its release in October 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

Meta
Own leaderboard Not tested since Oct 2024
Llama 3.2 11B 41.4%

Source: Official xMMMU leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) Aya Vision 32B Llama 3.2 11B Cambrian-1-8B 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) Aya Vision 32B Cambrian-1-8B

Sources: Official xMMMU leaderboard and labs’ model cards

Labs report it only in system cards

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.

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.

DateDocumentScoreAs the document writes itSection
May 13, 2025 Aya Vision 32B technical report 45.1 xMMMU, Aya Vision languages (Table 4) Capabilities

More in Vision

All 281 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. GPT-5.6 Sol 84.6%
  • 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%
  • MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy. Qwen 3.8 Max 85.9%
  • 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%