Model results

MMMU-Pro

Choose a shared evaluation protocol to compare model scores. Versions, sources and configurations stay separate.

Atlas snapshot captured 10/19/2026, 10:20:28 AM UTC. This is an import timestamp, not the evaluation date. Coverage may be incomplete.

Reviewed recommendation and limits

Prefer Pro to original MMMU when testing whether a model actually needs to understand the image.

MMMU-Pro standard and vision-only settings remain separate. Imported groups below may use different protocols; this review does not certify each run. Original MMMU test answers are public; inspect the Pro release separately. The project leaderboard is dated September 2025. Keep standard and vision-only settings separate.

Publisher results MMMU-Pro paper

← All evaluation groups

MMMU Pro

Source: Artificial Analysis
Version: Not reported
Measurement date: Not reported
Metric: % · Higher is better
Evaluation scope: Multimodal expert reasoning evaluation; observed 2026-10-19; result date not supplied

Exact score ties share a position. Positions do not establish statistical significance. Missing metadata is shown as unreported; matching imported metadata is not independent verification of the protocol.

#1Claude Opus 5.5
anthropic/claude-opus-5-5
87.7%
View modelEvidence independent
#2GPT-6 Astra
openai/gpt-6-astra-20260903
86.9%
View modelEvidence independent
#3GPT-6.1 Sol
openai/gpt-6.1-sol-20260929
86%
View modelEvidence independent
#4Gemini 3.8 Flash
google/gemini-3.8-flash-20260902
85.6%
View modelEvidence independent
#5Gemini 3.7 Flash
google/gemini-3.7-flash-20260813
85.5%
View modelEvidence independent
#6Claude Opus 5
anthropic/claude-opus-5
84.7%
View modelEvidence independent
#7GPT-5.6 Sol
openai/gpt-5.6-sol-20260709
83.4%
View modelEvidence independent
#8GPT-6 Sol
openai/gpt-6-sol-20260922
82.9%
View modelEvidence independent
#9Qwen3.8 Max (0902)
qwen/qwen3.8-max-20260902
82.8%
View modelEvidence independent
#10Muse Spark Open
meta/muse-spark-open-20261015
80.9%
View modelEvidence independent
#11GPT-5.6 Terra
openai/gpt-5.6-terra-20260709
80.7%
View modelEvidence independent
#12DeepSeek V4.1 Flash 1017
deepseek/deepseek-v4.1-flash-20261017
80.6%
View modelEvidence independent
#13Kimi K3
moonshotai/kimi-k3-20260715
80.5%
View modelEvidence independent
#14GPT-6 Luna
openai/gpt-6-luna-20260922
79.7%
View modelEvidence independent
#15Gemini 3.5 Flash Lite
google/gemini-3.5-flash-lite-20260721
79%
View modelEvidence independent
#16GPT-5.6 Luna
openai/gpt-5.6-luna-20260709
78.6%
View modelEvidence independent
#16MiniMax M3
minimax/minimax-m3-20260531
78.6%
View modelEvidence independent
#18DeepSeek V4.1 Flash
deepseek/deepseek-v4.1-flash-20260910
77%
View modelEvidence independent
#19Mistral Large 4
mistralai/mistral-large-4-0-20261006
76.4%
View modelEvidence independent
#19Step 5 Preview
stepfun/step-5-preview-20261008
76.4%
View modelEvidence independent
#21Qwen3.8 27B
qwen/qwen3.8-27b-20260814
76.3%
View modelEvidence independent
#22MiMo-V2.5
xiaomi/mimo-v2.5-20260422
75.4%
View modelEvidence independent
#23Inkling
thinkingmachines/inkling-20260715
73.5%
View modelEvidence independent
#24Mistral Medium 3
mistralai/mistral-medium-3
64.9%
View modelEvidence independent
#24Mistral Medium 3.1
mistralai/mistral-medium-3.1
64.9%
View modelEvidence independent
#24Mistral Medium 3.5
mistralai/mistral-medium-3.5-20260430
64.9%
View modelEvidence independent
#27Command A
cohere/command-a-03-2025
63.2%
View modelEvidence independent
#27Command A+
cohere/command-a-plus-05-2026
63.2%
View modelEvidence independent