The Reasoning Leaderboard shows how AI models able to handle complex logical and analytical tasks. It measures deep logical deduction, counterfactual reasoning, scientific problem-solving, and consistency across long, multi-turn interactions. These benchmarks are especially relevant for tasks that require more than simple information retrieval or pattern recognition.

Our composite score aggregates several leading reasoning benchmarks. Together, they highlight the models that perform best on complex reasoning, scientific analysis, and multi-step problem-solving.

Looking for the full picture? Reasoning accounts for 30% of the overall evaluation. Visit the Overall LLM Ranking to see how top reasoning models perform across coding, mathematics, and autonomous agent tasks.

Last updated: October 9, 2026
Rank Model Name Relative Quality Score
1 Claude Opus 5.5
Quality of Reasoning 97.52%

97.52
2 GPT-6 Astra
Quality of Reasoning 95.19%

95.19
3 Gemini 4 Argon
Quality of Reasoning 94.96%

94.96
4 Claude Fable 5.1
Quality of Reasoning 93.63%

93.63
5 Claude Opus 5
Quality of Reasoning 93.59%

93.59
6 Claude Fable 5
Quality of Reasoning 93.25%

93.25
7 Claude Sonnet 5.5
Quality of Reasoning 92.89%

92.89
8 GPT-5.6 Sol
Quality of Reasoning 91.86%

91.86
9 GPT-6.1 Sol
Quality of Reasoning 91.83%

91.83
10 Kimi K3
Quality of Reasoning 91.03%

91.03
11 Muse Spark 1.3
Quality of Reasoning 90.66%

90.66
12 Claude Opus 4.8
Quality of Reasoning 90.23%

90.23
13 Gemini 3.8 Flash
Quality of Reasoning 89.39%

89.39
14 Muse Spark 1.1
Quality of Reasoning 88.73%

88.73
15 GPT-5.5
Quality of Reasoning 88.64%

88.64
16 Qwen3.8
Quality of Reasoning 88.61%

88.61
17 Gemini 3.7 Flash
Quality of Reasoning 88.46%

88.46
18 GLM-5.3
Quality of Reasoning 87.57%

87.57
19 GPT-5.6 Terra
Quality of Reasoning 87.56%

87.56
20 Gemini 3.1 Pro
Quality of Reasoning 87.26%

87.26
21 Hy4
Quality of Reasoning 87.15%

87.15
22 Grok 4.6
Quality of Reasoning 87.13%

87.13
23 Claude Opus 4.7
Quality of Reasoning 87.08%

87.08
24 Claude Opus 4.6
Quality of Reasoning 86.77%

86.77
25 DeepSeek-V4-Pro
Quality of Reasoning 86.52%

86.52
26 Claude Sonnet 5
Quality of Reasoning 86.14%

86.14
27 GPT-5.4
Quality of Reasoning 86.04%

86.04
28 GPT-6 Sol
Quality of Reasoning 85.91%

85.91
29 Qwen3.7
Quality of Reasoning 85.17%

85.17
30 Grok 4.5
Quality of Reasoning 85.14%

85.14
31 Qwen3.8-Flash-Next
Quality of Reasoning 85.00%

85.00
32 Gemini 3.5 Flash
Quality of Reasoning 84.27%

84.27
33 DeepSeek-V4.1-Flash
Quality of Reasoning 83.93%

83.93
34 MiMo-V2.6-Pro
Quality of Reasoning 83.89%

83.89
35 GLM-5.3-Flash
Quality of Reasoning 83.80%

83.80
36 Gemini 3.6 Flash
Quality of Reasoning 83.66%

83.66
37 Seed 2.1 Pro
Quality of Reasoning 83.56%

83.56
38 Muse Spark 1.2
Quality of Reasoning 83.54%

83.54
39 GPT-5.6 Luna
Quality of Reasoning 83.14%

83.14
40 GLM-5.2
Quality of Reasoning 82.82%

82.82
41 GPT-5.2
Quality of Reasoning 82.65%

82.65
42 DeepSeek-V4-Flash
Quality of Reasoning 82.45%

82.45
43 Claude Haiku 5.5
Quality of Reasoning 81.99%

81.99
44 Kimi K2.6
Quality of Reasoning 81.55%

81.55
45 Step 5
Quality of Reasoning 81.51%

81.51
46 Muse Spark
Quality of Reasoning 81.24%

81.24
47 Gemini 3 Pro
Quality of Reasoning 80.86%

80.86
48 Grok 4.7
Quality of Reasoning 80.60%

80.60
49 Qwen3.8-27B
Quality of Reasoning 80.48%

80.48
50 MiniMax M3
Quality of Reasoning 80.08%

80.08
51 Claude Sonnet 4.6
Quality of Reasoning 80.03%

80.03
52 Qwen3.7-Plus
Quality of Reasoning 79.83%

79.83
53 Grok Build 0.1
Quality of Reasoning 79.02%

79.02
54 Claude Opus 4.5
Quality of Reasoning 78.89%

78.89
55 Gemini 3 Flash
Quality of Reasoning 78.63%

78.63
56 Qwen3.6
Quality of Reasoning 78.58%

78.58
57 Mistral Large 4
Quality of Reasoning 78.54%

78.54
58 Seed 2.0 Pro
Quality of Reasoning 78.54%

78.54
59 Hy3
Quality of Reasoning 78.22%

78.22
60 GPT-6 Luna
Quality of Reasoning 78.21%

78.21
61 Ling 3.1 Flash
Quality of Reasoning 76.32%

76.32
62 Inkling
Quality of Reasoning 76.29%

76.29
63 Qwen3.6 Plus
Quality of Reasoning 76.28%

76.28
64 GPT-5.3 Codex
Quality of Reasoning 76.22%

76.22
65 MiMo-V2.6-Flash
Quality of Reasoning 76.16%

76.16
66 GLM-5.1
Quality of Reasoning 76.13%

76.13
67 Kimi K2.5
Quality of Reasoning 75.43%

75.43
68 Kimi K2.7 Code
Quality of Reasoning 75.40%

75.40
69 Qwen3.5-397B-A17B
Quality of Reasoning 75.03%

75.03
70 MiMo-V2-Pro
Quality of Reasoning 74.98%

74.98
71 LongCat-Flash
Quality of Reasoning 74.95%

74.95
72 GPT-5.2 Codex
Quality of Reasoning 74.63%

74.63
73 Inkling-Small
Quality of Reasoning 74.29%

74.29
74 GPT-5.1
Quality of Reasoning 73.96%

73.96
75 GLM-5
Quality of Reasoning 73.80%

73.80
76 GPT-5.4 mini
Quality of Reasoning 72.93%

72.93
77 MiMo-V2.5
Quality of Reasoning 72.14%

72.14
78 GLM-4.7
Quality of Reasoning 72.07%

72.07
79 MiniMax M2.7
Quality of Reasoning 71.99%

71.99
80 MiMo-V2.5-Pro
Quality of Reasoning 71.84%

71.84
81 Qwen3
Quality of Reasoning 71.84%

71.84
82 GPT-5
Quality of Reasoning 71.80%

71.80
83 Grok 4.3
Quality of Reasoning 71.69%

71.69
84 Gemma 4 31B
Quality of Reasoning 71.29%

71.29
85 KAT-Coder-Pro V2
Quality of Reasoning 71.10%

71.10
86 GPT-5.4 nano
Quality of Reasoning 70.76%

70.76
87 Muse Glimmer
Quality of Reasoning 70.38%

70.38
88 Qwen3.5-122B-A10B
Quality of Reasoning 70.28%

70.28
89 GPT-5.5 Instant
Quality of Reasoning 70.09%

70.09
90 Grok-4
Quality of Reasoning 69.94%

69.94
91 ChatGPT-4o
Quality of Reasoning 69.85%

69.85
92 Kimi K2
Quality of Reasoning 69.83%

69.83
93 Qwen3.5-27B
Quality of Reasoning 69.35%

69.35
94 DeepSeek-V3.2
Quality of Reasoning 69.26%

69.26
95 Step-3.5-Flash
Quality of Reasoning 69.04%

69.04
96 Grok-4.20
Quality of Reasoning 68.47%

68.47
97 MiniMax M2.5
Quality of Reasoning 68.37%

68.37
98 Claude Sonnet 4.5
Quality of Reasoning 68.31%

68.31
99 Grok 4.1 Fast
Quality of Reasoning 68.14%

68.14
100 Solar Pro 4
Quality of Reasoning 67.92%

67.92