CtrlAltDebrief

MMMU-Pro benchmark

137 AI models ranked on MMMU-Pro. Each row shows the model’s score next to its cheapest API price per 1M tokens where it has one, so you can weigh quality against cost.

Top model today

Gemini 3.5 Flash (Google) leads at 84%, from $1.50 per 1M input.

Data via arXiv:2409.02813 →

Models ranked

137

on this benchmark

Top score

84%

current leader

Type

Single eval

benchmark

Updated

15 July 2026

last source fetch

Cost vs MMMU-Pro

85 priced models
Most attractive quadrant (cheap + high)
= reasoning model (extended thinking)
0%20%40%60%80%$0.1$0.3$1$3$10COST / 1M INPUT TOKENS (log) →MMMU-PRO SCORE ↑Gemini 3.5 FlashGPT-5.6 SolGemini 3.1 Pro PreviewGPT-5.5GPT-5.6 TerraGrok 4.5Gemini 3 ProClaude Opus 4.7MiniMax M3Gemini 3.1 Flash Lite PreviewQwen3.6 35B A3BGemma 4 26B A4bGemini 2.5 Flash Lite Preview 09 2025GPT-5 NanoGemma 4 E2bGemma 3 4BLlama 3.2 11B Vision Instruct

Each labelled name links to that model’s page. Curated to the cost-efficiency frontier, top scorers and cheapest so labels stay readable; the full field is in the leaderboard table below (and via “select all”). Reasoning models (extended “thinking” before answering) carry a ring around the dot.

Models ranked on MMMU-Pro

highest score first · 137 models
Maker
#
1Gemini 3.5 FlashGoogle84%$1.50$9.00
2GPT-5.6 SolOpenAI83%$5.00$30.00
3Gemini 3.1 Pro PreviewGoogle82%$2.00$12.00
4GPT-5.5OpenAI81%medium$5.00$30.00
5GPT-5.6 TerraOpenAI81%$2.50$15.00
6Muse SparkMeta81%n/an/a
7Qwen3.7 PlusAlibaba80%n/an/a
8Grok 4.5xAI80%$2.00$6.00
9Gemini 3 ProGoogle80%$2.00$12.00
10Gemini 3 FlashGoogle80%reasoning: truen/an/a
11Kimi K2.6Moonshot79%n/an/a
12Claude Opus 4.7Anthropic79%$5.00$25.00
13MiniMax M3MiniMax79%$0.30$1.20
14GPT-5.6 LunaOpenAI79%$1.00$6.00
15GPT-5.3 CodexOpenAI79%n/an/a
16GPT-5.4OpenAI78%$2.50$15.00
17Grok 4.3xAI78%$1.25$2.50
18Qwen3.6 PlusAlibaba78%$0.33$1.95
19Claude Sonnet 5Anthropic77%$2.00$10.00
20Qwen3.5 397B A17BAlibaba77%$0.60$3.60
21Grok Build 0.1 0616Spacexai77%n/an/a
22GPT-5.2-CodexOpenAI76%$1.75$14.00
23Gemini 3.1 Flash Lite PreviewGoogle76%$0.25$1.50
24GPT-5.1OpenAI75%$1.25$10.00
25Claude Opus 4.6Anthropic75%reasoning: adaptive$5.00$25.00
26MiMo-V2.5Xiaomi75%n/an/a
27Kimi K2.5Moonshot75%$0.50$2.80
28Step 3.7 FlashStepfun75%n/an/a
29Qwen3.6 35B A3BAlibaba75%$0.14$0.45
30Qwen3.5-27BAlibaba75%$0.30$2.40
31Qwen3.5-122B-A10BAlibaba75%$0.25$1.75
32Gemini 2.5 ProGoogle75%$1.25$10.00
33Qwen3.6 27BAlibaba75%$0.15$0.50
34GPT-5.2OpenAI75%medium$1.75$14.00
35Grok 4.20 0309 v2Spacexai75%n/an/a
36GPT-5OpenAI74%medium$1.25$10.00
37Claude Opus 4.5Anthropic74%reasoning: true$5.00$25.00
38MiMo-V2-Omni-0327Xiaomi74%n/an/a
39GPT-5 CodexOpenAI74%$1.25$10.00
40Gemma 4 31BGoogle73%$0.14$0.40
41Claude Sonnet 4.6Anthropic73%reasoning: adaptive$3.00$15.00
42GPT-5.4 MiniOpenAI73%$0.75$4.50
43Grok 4.20 0309Spacexai73%n/an/a
44Gemini 2.5 Flash Preview 09 2025Google73%reasoning: true$0.30$2.50
45GLM 5V TurboZhipu73%n/an/a
46Qwen3.5-35B-A3BAlibaba73%$0.25$2.00
47GPT-5.1 CodexOpenAI72%n/an/a
48Qwen3.5 Omni PlusAlibaba71%n/an/a
49GPT-5 MiniOpenAI70%$0.25$2.00
50o3OpenAI70%$2.00$8.00
1–50 of 137
Page 1 of 3
Prices are on-demand list rates. Enterprise commitments can be materially lower. Cost per task uses best-effort token usage.

Frequently asked questions

What is the MMMU-Pro benchmark?

A more robust version of the MMMU multimodal benchmark testing reasoning over combined images and text across many disciplines. It filters out text-only-answerable questions, expands the answer options, and adds a vision-only mode where the question is embedded in the image. Scored as multiple-choice accuracy.

Which AI model scores highest on MMMU-Pro?

Gemini 3.5 Flash (Google) leads with 84%, from $1.50 per 1M input tokens.

How many models are ranked on MMMU-Pro?

137 models carry a MMMU-Pro score. Scores come from Artificial Analysis; where a model has a live API price it is verified against LiteLLM and OpenRouter, and models with no current price are listed without one.

Other benchmarks

compare the same models on a different eval
Intelligence Indexcomposite indexCoding Indexcomposite indexAgentic Indexcomposite indexGPQA DiamondleaderboardHumanity's Last ExamleaderboardMMLU-ProleaderboardLiveCodeBenchleaderboardSciCodeleaderboardTerminalBench HardleaderboardTerminalBench v2.1leaderboardτ²-benchleaderboardτ-bench BankingleaderboardAPEX AgentsleaderboardIT-Bench SREleaderboardGDPvalleaderboardCritPtleaderboardOmniscienceleaderboardIFBenchleaderboardLong-Context Reasoningcomposite indexSWE-bench VerifiedleaderboardSWE-bench ProleaderboardAIME 2026leaderboardARC-AGI-2leaderboardAA-Briefcaseleaderboard
Benchmark scores via Artificial Analysis · pricing verified against LiteLLM + OpenRouter · refreshed daily · updated 15 July 2026

Every weekday

AI moves fast. Here's your debrief.

News, analysis, tools, and more.

For people who build with AI