MMMU leaderboard

College-level questions that need both an image and text: charts, diagrams, tables and figures. Claude Fable 5.1 leads with 90.64%.

Models
81
Leader
Claude Fable 5.1
Source
Vals AI
As of
Oct 24, 2026
  1. 1Claude Fable 5.190.6%
  2. 2Claude Opus 589.9%
  3. 3Claude Fable 589.3%
  4. 4Gemini 3.8 Flash89.1%
  5. 5Gemini 3.7 Flash89.0%
  6. 6GPT-5.6 Sol88.8%
  7. 7Gemini 3.6 Flash88.4%
  8. 8Gemini 3.5 Flash88.3%
  9. 9GPT-5.588.3%
  10. 10Gemini 3.1 Pro Preview88.2%
MMMU, accuracy against cost per test
Open weightsProprietaryBest for the price
10%33%55%78%100%$.1$1$10Cost per test, log scale →Claude Fable 5.1Gemini 3.8 FlashGemini 3 FlashGLM-5.3-FlashGPT-5 nano

Full leaderboard

RankModelLabAccuracyCost / testHarnessPrice in / out
1Claude Fable 5.1Anthropic90.64%$10 / $50—$10 / $50
2Claude Opus 5Anthropic89.88%$5 / $25—$5 / $25
3Claude Fable 5Anthropic89.31%$10 / $50—$10 / $50
4Gemini 3.8 FlashGoogle89.08%$1.5 / $7.5—$0.82 / $4.13
5Gemini 3.7 FlashGoogle88.96%$1.5 / $7.5—$0.75 / $3.75
6GPT-5.6 SolOpenAI88.84%$4 / $20—$4 / $20
7Gemini 3.6 FlashGoogle88.38%$1.5 / $7.5—$0.75 / $3.75
8Gemini 3.5 FlashGoogle88.27%$1.5 / $9—$1.50 / $9
9GPT-5.5OpenAI88.27%$5 / $30—$5 / $30
10Gemini 3.1 Pro PreviewGoogle88.21%$2 / $12—$2 / $12
11Kimi K3Moonshot AI88.15%$3 / $15—$0.24 / $11
12Qwen3.8-MaxAlibaba88.03%$2 / $6—$2 / $6
13Gemini 3 FlashGoogle87.63%$0.5 / $3—$0.50 / $3
14Gemini 3 ProGoogle87.51%$2 / $12——
15GPT-5.4OpenAI87.51%$2.5 / $15—$2.50 / $15
16Muse SparkMeta87.40%N/A——
17GPT-5.2OpenAI86.67%$1.75 / $14—$1.75 / $14
18Muse Spark 1.1Meta86.59%$1.25 / $4.25—$1.25 / $4.25
19Claude Opus 4.8Anthropic86.59%$5 / $25—$5 / $25
20GPT-5.6 TerraOpenAI86.47%$2 / $12—$2 / $12
21Kimi K2.6Moonshot AI86.30%$0.95 / $4—$0.65 / $3.41
22Muse Spark 1.2Meta86.13%$1.25 / $4.25—$1.25 / $4.25
23GLM-5.3-FlashZ.ai86.01%$0.075 / $0.25—$0.15 / $0.50
24Claude Opus 4.7Anthropic85.55%$5 / $25—$5 / $25
25GPT-5.6 LunaOpenAI85.03%$0.2 / $1.2—$0.20 / $1.20
26Kimi K2.5Moonshot AI84.33%$0.6 / $3—$0.50 / $2.50
27Qwen3.6-PlusAlibaba84.16%$0.5 / $3—$0.33 / $1.95
28Qwen3.8-27BAlibaba83.93%$0.5 / $3—$0.20 / $2.50
29Claude Opus 4.6Anthropic83.87%$5 / $25—$5 / $25
30Gemini 3.5 Flash-LiteGoogle83.58%$0.3 / $2.5—$0.30 / $2.50
31Claude Sonnet 4.6Anthropic83.58%$3 / $15—$3 / $15
32Grok 4.20xAI83.47%$2 / $6—$1.25 / $2.50
33GPT-5.1OpenAI83.18%$1.25 / $10—$1.25 / $10
34Grok 4.3xAI83.06%$1.25 / $2.5—$1.25 / $2.50
35Claude Sonnet 5Anthropic83.01%$2 / $10—$2 / $10
37Gemini 3.1 Flash-Lite PreviewGoogle82.49%$0.25 / $1.5—$0.25 / $1.50
38Qwen3.5-FlashAlibaba81.91%$0.1 / $0.4—$0.065 / $0.26
39GPT-5OpenAI81.50%$1.25 / $10—$1.25 / $10
40Gemini 2.5 ProGoogle81.34%$1.25 / $10—$1.25 / $10
41MiniMax-M3MiniMax81.16%$0.6 / $2.4—$0.30 / $1.20
42Claude Opus 4.5Anthropic81.10%$5 / $25—$5 / $25
44o3OpenAI80.42%$2 / $8—$2 / $8
45MiMo-V2.5Xiaomi80.00%$0.14 / $0.28—$0.12 / $0.24
46o4-miniOpenAI79.67%$1.1 / $4.4—$1.10 / $4.40
48Claude Sonnet 4.5Anthropic79.31%$3 / $15—$3 / $15
49GPT-5.4 miniOpenAI79.25%$0.75 / $4.5—$0.75 / $4.50
50GPT-5 miniOpenAI78.91%$0.25 / $2—$0.25 / $2
51InklingThinky78.50%$1 / $4.05—$1 / $4.05
53o1OpenAI77.41%$15 / $60—$15 / $60
54Grok 4xAI76.27%$3 / $15——
58Claude Opus 4.1Anthropic73.72%$15 / $75—$15 / $75
59GPT-5.4 nanoOpenAI73.58%$0.2 / $1.25—$0.20 / $1.25
60Claude Opus 4Anthropic73.31%$15 / $75——
62Grok 4.1 FastxAI72.66%$0.2 / $0.5——
63Gemini 2.5 Flash PreviewGoogle72.54%$0.1 / $0.4——
64Claude Sonnet 4Anthropic72.39%$3 / $15——
65GPT-4.1OpenAI72.39%$2 / $8—$2 / $8
67Llama 4 MaverickFireworks AI71.69%$0.22 / $0.88—$0.19 / $0.65
68Claude 3.7 SonnetAnthropic71.52%$3 / $15——
69GPT-5 nanoOpenAI70.94%$0.05 / $0.4—$0.05 / $0.40
70GPT-4.1 miniOpenAI70.54%$0.4 / $1.6—$0.40 / $1.60
71Gemini 2.0 FlashGoogle69.79%$0.1 / $0.4——
72Claude 3.5 SonnetAnthropic68.80%$3 / $15——
73Gemini 2.5 Flash Preview 4/17Google68.00%$0.3 / $2.5——
74Mistral LargeMistral AI66.19%$0.5 / $1.5—$0.50 / $1.50
75Gemini 1.5 ProGoogle65.51%$1.25 / $5——
77Magistral MediumMistral AI64.57%$2 / $5——
79Grok 4.1 Fast Non-ReasoningxAI63.70%$0.2 / $0.5——
80Grok 4 FastxAI63.41%$0.2 / $0.5——
81Mistral MediumMistral AI62.97%$0.4 / $2——
82GPT-4oOpenAI62.16%$2.5 / $10—$2.50 / $10
83Grok 4.5xAI61.79%$2 / $6—$2 / $6
84Magistral Small 1.2Mistral AI60.08%$0.075 / $0.3——
85Llama 4 ScoutMeta58.75%$0.18 / $0.59—$0.10 / $0.30
86Grok 2 VisionxAI57.25%$2 / $10——
87Gemini 1.5 FlashGoogle57.19%$0.075 / $0.3——
88GPT-4o miniOpenAI56.56%$0.15 / $0.6—$0.15 / $0.60
89GPT-4.1 nanoOpenAI55.05%$0.1 / $0.4—$0.10 / $0.40
91Claude Haiku 4.5Anthropic46.07%$1 / $5—$1 / $5
92Llama 3.2 VisionMeta38.82%$0.18 / $0.18——
93Qwen 3.5 PlusAlibaba22.77%$0.4 / $2.4—$0.30 / $1.80

Related leaderboards

Sources

  • Vals AI, as of Oct 24, 2026
  • Provider list prices, as of Oct 24, 2026

Every number on this page comes from these sources; models without a published number show none. Download the data. Cite as “Nexus Explorer, nexus.inc/models”.