The Math Ranking evaluates the mathematical rigor and quantitative reasoning capabilities of large language models. Rather than focusing on simple pattern recognition, these benchmarks test performance on formal proofs, competition-level Olympiad problems, multi-step algebra, and discrete mathematics, where accuracy and logical consistency are critical.

By aggregating standardized scores from various sources this leaderboard highlights the models that perform best in mathematical problem-solving, multi-step reasoning, and numerical accuracy.

Looking for a balanced model across all domains? Mathematical performance contributes 20% to the composite score. Visit the Overall LLM Ranking to see how top math models compare in coding, agentic capabilities, and broader logical reasoning.

Last updated: October 19, 2026
Rank Model Name Relative Quality Score
1 Gemini 4 Argon
Quality of Math 97.26%

97.26
2 Claude Fable 5.1
Quality of Math 96.87%

96.87
3 Claude Opus 5
Quality of Math 95.07%

95.07
4 Claude Fable 5
Quality of Math 94.61%

94.61
5 Claude Opus 5.5
Quality of Math 94.32%

94.32
6 MiMo-V2.6-Pro
Quality of Math 93.57%

93.57
7 Muse Spark 1.2
Quality of Math 92.95%

92.95
8 Muse Spark 1.1
Quality of Math 92.31%

92.31
9 Qwen3.6
Quality of Math 91.69%

91.69
10 Claude Sonnet 5.5
Quality of Math 91.48%

91.48
11 GPT-5.5
Quality of Math 90.74%

90.74
12 MiMo-V2.6-Flash
Quality of Math 90.60%

90.60
13 Muse Spark 1.3
Quality of Math 90.52%

90.52
14 Gemini 3.8 Flash
Quality of Math 90.27%

90.27
15 GPT-6 Sol
Quality of Math 90.25%

90.25
16 Claude Opus 4.8
Quality of Math 89.98%

89.98
17 GPT-6.1 Sol
Quality of Math 89.96%

89.96
18 Kimi K3
Quality of Math 89.82%

89.82
19 Gemini 3.7 Flash
Quality of Math 89.73%

89.73
20 Muse Spark
Quality of Math 89.10%

89.10
21 Step 5
Quality of Math 89.03%

89.03
22 Claude Opus 4.6
Quality of Math 88.94%

88.94
23 GLM-5.3
Quality of Math 88.79%

88.79
24 GPT-6 Astra
Quality of Math 88.58%

88.58
25 GPT-5.6 Sol
Quality of Math 88.53%

88.53
26 Claude Opus 4.7
Quality of Math 88.33%

88.33
27 Qwen3.7
Quality of Math 88.27%

88.27
28 Gemini 3 Pro
Quality of Math 88.00%

88.00
29 Muse Wonder
Quality of Math 87.62%

87.62
30 MiMo-V2-Pro
Quality of Math 87.62%

87.62
31 GPT-5.5 Pro
Quality of Math 87.50%

87.50
32 GPT-5.4
Quality of Math 87.42%

87.42
33 Muse Glimmer
Quality of Math 87.15%

87.15
34 Kimi K2.5
Quality of Math 87.06%

87.06
35 Gemini 3.1 Pro
Quality of Math 86.95%

86.95
36 GLM-5
Quality of Math 86.68%

86.68
37 Grok-4.1
Quality of Math 86.36%

86.36
38 Hy3
Quality of Math 86.35%

86.35
39 GLM-5.2
Quality of Math 86.11%

86.11
40 Hy4
Quality of Math 86.03%

86.03
41 GPT-6 Luna
Quality of Math 85.85%

85.85
42 MiMo-V2.5
Quality of Math 85.74%

85.74
43 GLM-5V-Turbo
Quality of Math 85.74%

85.74
44 GPT-5.6 Terra
Quality of Math 85.43%

85.43
45 Qwen3.8
Quality of Math 85.40%

85.40
46 DeepSeek-V4-Pro
Quality of Math 85.25%

85.25
47 MiMo-V2-Omni
Quality of Math 85.11%

85.11
48 Gemini 3.6 Flash
Quality of Math 84.80%

84.80
49 Claude Sonnet 4.6
Quality of Math 84.78%

84.78
50 DeepSeek-V4.1-Flash
Quality of Math 84.53%

84.53
51 Claude Sonnet 5
Quality of Math 84.27%

84.27
52 GPT-5.2
Quality of Math 83.91%

83.91
53 GLM-5.3-Flash
Quality of Math 83.87%

83.87
54 GPT-5.2 Pro
Quality of Math 83.78%

83.78
55 Kimi K2
Quality of Math 83.53%

83.53
56 MiniMax M2.7
Quality of Math 83.07%

83.07
57 Qwen3.5-122B-A10B
Quality of Math 82.98%

82.98
58 GPT-5.6 Luna
Quality of Math 82.40%

82.40
59 Kimi K2.6
Quality of Math 82.36%

82.36
60 Qwen3.5-27B
Quality of Math 82.12%

82.12
61 GLM-4.7
Quality of Math 82.04%

82.04
62 GPT-5.1
Quality of Math 81.96%

81.96
63 Seed 2.0 Pro
Quality of Math 81.88%

81.88
64 DeepSeek-V4-Flash
Quality of Math 81.67%

81.67
65 Grok 4.6
Quality of Math 81.26%

81.26
66 Qwen3.8-27B
Quality of Math 81.25%

81.25
67 Gemma 4 31B
Quality of Math 81.25%

81.25
68 Gemini 3 Flash
Quality of Math 81.14%

81.14
69 Grok 4.5
Quality of Math 81.06%

81.06
70 Mistral Large 4
Quality of Math 81.02%

81.02
71 Gemini 3.5 Flash
Quality of Math 80.91%

80.91
72 ChatGPT-4o
Quality of Math 80.41%

80.41
73 Step-3.5-Flash
Quality of Math 79.84%

79.84
74 ERNIE 5.0
Quality of Math 79.76%

79.76
75 DeepSeek-V3.2
Quality of Math 79.60%

79.60
76 Solar Pro 4
Quality of Math 79.53%

79.53
77 DeepSeek V3.1 Terminus
Quality of Math 79.31%

79.31
78 Grok-4
Quality of Math 79.29%

79.29
79 DeepSeek-V3.2-Speciale
Quality of Math 79.28%

79.28
80 LongCat-Flash
Quality of Math 79.21%

79.21
81 Mistral Medium 3.1
Quality of Math 79.00%

79.00
82 MiMo-V2.5-Pro
Quality of Math 78.89%

78.89
83 MiniMax M2.5
Quality of Math 78.68%

78.68
84 Qwen3.7-Plus
Quality of Math 78.56%

78.56
85 Qwen3 235B A22B
Quality of Math 78.51%

78.51
86 Qwen3.5-35B-A3B
Quality of Math 78.27%

78.27
87 Qwen3.6 Plus
Quality of Math 78.24%

78.24
88 Gemma 4 26B-A4B
Quality of Math 77.95%

77.95
89 Qwen3.8-Flash-Next
Quality of Math 77.24%

77.24
90 Trinity Large
Quality of Math 77.12%

77.12
91 Qwen3.8 Flash
Quality of Math 77.08%

77.08
92 Qwen3 VL 235B A22B
Quality of Math 76.62%

76.62
93 Grok 4.7
Quality of Math 76.53%

76.53
94 Inkling-Small
Quality of Math 76.33%

76.33
95 Qwen3-Coder 480B A35B
Quality of Math 76.02%

76.02
96 Grok-4.20
Quality of Math 75.95%

75.95
97 Grok 4 Fast
Quality of Math 75.92%

75.92
98 GPT-5
Quality of Math 75.88%

75.88
99 GLM-5.1
Quality of Math 75.62%

75.62
100 INTELLECT-3
Quality of Math 75.55%

75.55