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: November 5, 2026
Rank Model Name Relative Quality Score
1 Claude Opus 5.5
Quality of Reasoning 97.99%

97.99
2 Gemini 4 Carbon
Quality of Reasoning 97.17%

97.17
3 GPT-6 Astra
Quality of Reasoning 96.29%

96.29
4 Gemini 4 Argon
Quality of Reasoning 95.55%

95.55
5 Claude Sonnet 5.5
Quality of Reasoning 95.44%

95.44
6 Kimi K3.1 Pro
Quality of Reasoning 94.91%

94.91
7 GPT-6.1 Sol
Quality of Reasoning 93.74%

93.74
8 Claude Opus 5
Quality of Reasoning 93.38%

93.38
9 Claude Fable 5.1
Quality of Reasoning 92.97%

92.97
10 Kimi K3
Quality of Reasoning 92.93%

92.93
11 Claude Fable 5
Quality of Reasoning 92.74%

92.74
12 GPT-5.6 Sol
Quality of Reasoning 92.27%

92.27
13 Muse Spark 1.3
Quality of Reasoning 92.22%

92.22
14 Claude Opus 4.8
Quality of Reasoning 90.96%

90.96
15 Qwen3.8
Quality of Reasoning 90.95%

90.95
16 DeepSeek-V4.1-Flash
Quality of Reasoning 90.43%

90.43
17 Muse Spark 1.1
Quality of Reasoning 90.33%

90.33
18 GLM-5.3
Quality of Reasoning 90.29%

90.29
19 GPT-5.6 Terra
Quality of Reasoning 90.03%

90.03
20 Gemini 3.8 Flash
Quality of Reasoning 89.87%

89.87
21 GPT-6.1 Terra
Quality of Reasoning 89.80%

89.80
22 Grok 4.6
Quality of Reasoning 89.35%

89.35
23 Step 5
Quality of Reasoning 89.32%

89.32
24 Claude Opus 4.7
Quality of Reasoning 89.15%

89.15
25 Muse Wonder
Quality of Reasoning 89.06%

89.06
26 Claude Opus 4.6
Quality of Reasoning 88.80%

88.80
27 GPT-5.5
Quality of Reasoning 87.94%

87.94
28 Claude Sonnet 5
Quality of Reasoning 87.87%

87.87
29 Qwen3.7
Quality of Reasoning 87.47%

87.47
30 DeepSeek-V4-Pro
Quality of Reasoning 87.41%

87.41
31 MiMo-V2.6-Pro
Quality of Reasoning 87.17%

87.17
32 Hy4
Quality of Reasoning 86.88%

86.88
33 GLM-5.3-Flash
Quality of Reasoning 86.74%

86.74
34 Gemini 3.1 Pro
Quality of Reasoning 86.63%

86.63
35 Gemini 3.7 Flash
Quality of Reasoning 86.49%

86.49
36 Muse Glimmer
Quality of Reasoning 86.23%

86.23
37 GPT-6 Sol
Quality of Reasoning 86.21%

86.21
38 GLM-5.2
Quality of Reasoning 85.99%

85.99
39 Claude Haiku 5.5
Quality of Reasoning 85.92%

85.92
40 Qwen3.8-Flash-Next
Quality of Reasoning 85.77%

85.77
41 Kimi K2.6
Quality of Reasoning 85.51%

85.51
42 Mistral Large 4
Quality of Reasoning 85.09%

85.09
43 Grok 4.5
Quality of Reasoning 84.81%

84.81
44 MiMo-V2.6-Flash
Quality of Reasoning 84.73%

84.73
45 GPT-5.4
Quality of Reasoning 84.70%

84.70
46 Gemini 3.5 Flash
Quality of Reasoning 84.52%

84.52
47 Muse Spark
Quality of Reasoning 84.24%

84.24
48 Qwen3.8-27B
Quality of Reasoning 84.23%

84.23
49 MiniMax M3.1 Flash
Quality of Reasoning 84.17%

84.17
50 Grok 4.7
Quality of Reasoning 83.95%

83.95
51 GPT-5.6 Luna
Quality of Reasoning 83.81%

83.81
52 Qwen3.7-Plus
Quality of Reasoning 83.80%

83.80
53 Claude Opus 4.5
Quality of Reasoning 83.76%

83.76
54 Hy3
Quality of Reasoning 83.74%

83.74
55 DeepSeek V3.1 Terminus
Quality of Reasoning 83.62%

83.62
56 GPT-5.2
Quality of Reasoning 83.09%

83.09
57 GPT-6 Luna
Quality of Reasoning 83.02%

83.02
58 Gemini 3.6 Flash
Quality of Reasoning 82.89%

82.89
59 Mistral Medium 3.1
Quality of Reasoning 82.61%

82.61
60 Claude Sonnet 4.6
Quality of Reasoning 82.31%

82.31
61 MiniMax M3
Quality of Reasoning 82.19%

82.19
62 GPT-5.5 Pro
Quality of Reasoning 82.04%

82.04
63 Gemini 3 Pro
Quality of Reasoning 81.85%

81.85
64 GLM-5.1
Quality of Reasoning 81.84%

81.84
65 Seed 2.0 Pro
Quality of Reasoning 81.84%

81.84
66 Muse Wonder Full
Quality of Reasoning 81.63%

81.63
67 Qwen3.6 Plus
Quality of Reasoning 81.57%

81.57
68 Beam 1.0
Quality of Reasoning 81.49%

81.49
69 Trinity Large
Quality of Reasoning 81.16%

81.16
70 Kimi K2.5
Quality of Reasoning 80.90%

80.90
71 DeepSeek-V4-Flash
Quality of Reasoning 80.71%

80.71
72 Ling 3.0 Flash Fin
Quality of Reasoning 80.39%

80.39
73 MiMo-V2-Pro
Quality of Reasoning 80.36%

80.36
74 Gemini 3 Flash
Quality of Reasoning 80.36%

80.36
75 Qwen3.5-Flash
Quality of Reasoning 80.00%

80.00
76 Muse Spark 1.2
Quality of Reasoning 79.87%

79.87
77 Qwen3.5-397B-A17B
Quality of Reasoning 79.83%

79.83
78 Inkling
Quality of Reasoning 79.81%

79.81
79 Muse Spark Open
Quality of Reasoning 79.79%

79.79
80 GLM-5
Quality of Reasoning 79.65%

79.65
81 GPT-5.1
Quality of Reasoning 78.52%

78.52
82 Nemotron 3.5 Super VL
Quality of Reasoning 78.44%

78.44
83 LongCat-Flash
Quality of Reasoning 78.42%

78.42
84 MiMo-V2.5
Quality of Reasoning 78.24%

78.24
85 Inkling-Small
Quality of Reasoning 77.90%

77.90
86 Claude Sonnet 4.5
Quality of Reasoning 76.89%

76.89
87 Kimi K2
Quality of Reasoning 76.65%

76.65
88 MiniMax M2.7
Quality of Reasoning 76.55%

76.55
89 Mistral Small 3.2
Quality of Reasoning 76.52%

76.52
90 GPT-5.5 Instant
Quality of Reasoning 76.46%

76.46
91 GLM-4.7
Quality of Reasoning 76.35%

76.35
92 ERNIE 5.0
Quality of Reasoning 76.10%

76.10
93 GPT-5.2 Pro
Quality of Reasoning 76.10%

76.10
94 Qwen3
Quality of Reasoning 75.93%

75.93
95 Step-3.5-Flash
Quality of Reasoning 75.82%

75.82
96 Gemma 4 31B
Quality of Reasoning 75.69%

75.69
97 MiniMax M2.5
Quality of Reasoning 75.60%

75.60
98 GPT-5.4 mini
Quality of Reasoning 75.56%

75.56
99 Claude Opus 4.1
Quality of Reasoning 75.41%

75.41
100 Qwen3.5-122B-A10B
Quality of Reasoning 75.37%

75.37