Artificial Analysis

Artificial Analysis MMMU-Pro

An Artificial Analysis MMMU-Pro score.

In index
Multimodal
Source ↗

Models101

Top result87.7%Claude Opus 5.5

Top-3 spread2.1 pts

Year2026

Result and price

Results

101 results

#ModelResult

01

Claude Opus 5.587.7%

02

GPT-6 Astra86.9%

03

Gemini 3.8 Flash85.6%

04

Gemini 3.7 Flash85.5%

05

Claude Opus 584.7%

06

Gemini 3.5 Flash84.3%

07

GPT-5.6 Sol83.4%

08

GPT-6 Sol83.3%

09

Gemini 3.6 Flash83.2%

10

Qwen3.8 Max Preview82.8%

11

Gemini 3.1 Pro82.4%

12

GPT-5.6 Terra80.7%

13MAKimi K380.5%

13

Muse Spark80.5%

13

Qwen3.7 Plus80.5%

16

Grok 4.580.4%

17

Gemini 3 Pro80.2%

18

GPT-5.579.9%

19

Qwen3.8-Flash-Next79.8%

20MAKimi K2.679.4%

21APApodex 1.179.2%

21APApodex 1.1 Mini79.2%

23

Gemini 3.5 Flash-Lite79.0%

23INLing 3.0 Flash VL79.0%

25

Claude Opus 4.7 (Adaptive)78.8%

26

GPT-5.6 Luna78.6%

26

MiniMax M378.6%

26

Gemini 3 Flash78.6%

29

GPT-5.3 Codex78.5%

30

GPT-5.478.4%

31

Grok 4.378.1%

32

Qwen3.6 Plus78.0%

33

Claude Sonnet 577.3%

34

DeepSeek V4.1 Flash77.0%

35STStep 5 Preview76.4%

35

Claude Opus 4.776.4%

37

Qwen3.8-27B76.3%

37

GPT-5.2-Codex76.3%

39

GPT-6 Luna75.5%

39

GPT-5.175.5%

41

Claude Opus 4.6 (Adaptive)75.4%

41MAKimi K2.5 (Reasoning)75.4%

41MAKimi K2.575.4%

44STStep 3.7 Flash75.3%

45

Qwen3.5-122B-A10B75.0%

45

Qwen3.5-27B75.0%

45

Qwen3.6-35B-A3B75.0%

48

Gemini 2.5 Pro74.9%

49

Qwen3.6-27B74.6%

50

GPT-5 (medium)74.3%

50

Muse Glimmer 30B74.3%

52

GPT-5 (high)74.2%

53

Claude Opus 4.5 Thinking74.0%

53TMInkling-Small74.0%

55TMInkling73.5%

56

Gemma 4 31B73.4%

57

GPT-5.4 mini73.3%

58

GLM-5V-Turbo72.8%

59

Qwen3.5-35B-A3B72.7%

60

Claude Opus 4.672.5%

60

GPT-5.1-Codex-Max72.5%

60

GPT-5.1-Codex72.5%

63

Claude Opus 4.571.2%

64

Claude Sonnet 4.670.6%

65

o370.1%

66

MiMo-V2-Omni69.9%

67

Gemma 4 12B69.7%

68

Gemma 4 26B A4B69.2%

69

Grok 468.8%

70

Claude 4.1 Opus Thinking67.9%

71

Gemini 2.5 Flash65.5%

72

GPT-5.4 nano65.4%

73

Mistral Medium 3.5 128B64.9%

74

Grok 4.1 Fast (Reasoning)63.3%

75COCommand A+63.2%

76

Claude 4 Sonnet62.4%

77

Llama 4 Maverick62.1%

78

Grok 4 Fast (Reasoning)61.8%

79

GPT-4.161.2%

80

Qwen3-Omni-30B-A3B-Thinking60.2%

81

GPT-4.1 mini58.7%

82

Mistral Small 456.8%

—

Mistral Small 4 (Reasoning)56.8%

83

Mistral Large 355.7%

84

Qwen3-Omni-30B-A3B-Instruct55.5%

85

Gemini 1.5 Pro55.0%

86

Nemotron 3 Nano Omni 30B A3B53.2%

87

Mistral Medium 353.0%

88

Llama 4 Scout52.9%

89

Qwen3.5 397B52.7%

89

Qwen3.5 397B (Reasoning)52.7%

91

Gemma 4 E4B51.4%

92

Grok 4.1 Fast48.4%

93

Gemma 3 27B48.0%

94

Gemma 4 E2B44.6%

95AMNova Pro44.3%

96

GPT-4o mini41.5%

97

GPT-4.1 nano40.1%

98

Claude 3 Haiku30.8%

99LILFM2.5-VL-1.6B-Extract26.5%

100

Phi-4 Multimodal Instruct14.5%