The Coding Leaderboard evaluates LLMs on their software engineering capabilities. From generating production-ready code across major programming languages to designing algorithms, debugging complex functions, and translating scientific concepts into executable scripts, this category measures a model’s ability to handle a broad range of coding and software development tasks.

Scores in this category are compiled from leading coding benchmarks highlighting models that perform well on code generation, debugging, problem-solving, and technical accuracy.

Need more than code generation? Coding performance accounts for 30% of our composite rating. Explore the Overall LLM Ranking to see which top coding models also excel at autonomous agent workflows, advanced mathematics, and complex reasoning.

Last updated: October 19, 2026
Rank Model Name Relative Quality Score
1 Claude Opus 5.5
Quality of Coding 99.35%

99.35
2 Claude Sonnet 5.5
Quality of Coding 94.94%

94.94
3 GPT-6 Astra
Quality of Coding 93.48%

93.48
4 Claude Fable 5.1
Quality of Coding 90.18%

90.18
5 GPT-6.1 Sol
Quality of Coding 90.13%

90.13
6 Muse Wonder Full
Quality of Coding 89.86%

89.86
7 Gemini 4 Argon
Quality of Coding 89.78%

89.78
8 ERNIE 5.0
Quality of Coding 89.64%

89.64
9 Claude Fable 5
Quality of Coding 89.38%

89.38
10 DeepSeek-V4.1-Flash
Quality of Coding 87.47%

87.47
11 Claude Opus 5
Quality of Coding 86.66%

86.66
12 ChatGPT-4o
Quality of Coding 86.39%

86.39
13 Muse Wonder
Quality of Coding 86.31%

86.31
14 Muse Spark 1.3
Quality of Coding 86.25%

86.25
15 GPT-5.6 Sol
Quality of Coding 86.12%

86.12
16 Qwen3 VL 235B A22B
Quality of Coding 85.80%

85.80
17 GPT-6 Sol
Quality of Coding 84.99%

84.99
18 GPT-5.5 Instant
Quality of Coding 84.93%

84.93
19 MiMo-V2.6-Pro
Quality of Coding 84.11%

84.11
20 GLM-5.3
Quality of Coding 83.55%

83.55
21 Grok 4.7
Quality of Coding 81.99%

81.99
22 Step 5
Quality of Coding 81.51%

81.51
23 Mistral Medium 3
Quality of Coding 81.21%

81.21
24 Grok 4.6
Quality of Coding 81.00%

81.00
25 Qwen3.8 Flash
Quality of Coding 80.71%

80.71
26 Kimi K3
Quality of Coding 80.62%

80.62
27 Gemini 3.8 Flash
Quality of Coding 79.88%

79.88
28 GPT-5.6 Terra
Quality of Coding 79.64%

79.64
29 Grok 4.5
Quality of Coding 79.59%

79.59
30 Claude Haiku 5.5
Quality of Coding 79.53%

79.53
31 Gemini 3.7 Flash
Quality of Coding 79.26%

79.26
32 Claude Opus 4.8
Quality of Coding 79.11%

79.11
33 GLM-4.6V
Quality of Coding 78.85%

78.85
34 INTELLECT-3
Quality of Coding 78.70%

78.70
35 MiMo-V2.6-Flash
Quality of Coding 78.64%

78.64
36 Qwen3.8
Quality of Coding 78.47%

78.47
37 DeepSeek-V4-Pro
Quality of Coding 78.30%

78.30
38 GPT-5.5
Quality of Coding 78.20%

78.20
39 Hy4
Quality of Coding 78.15%

78.15
40 Beam 1.0
Quality of Coding 78.06%

78.06
41 GLM-5.3-Flash
Quality of Coding 77.94%

77.94
42 Claude Opus 4.7
Quality of Coding 77.91%

77.91
43 Ling-flash-2.0
Quality of Coding 77.81%

77.81
44 Claude Sonnet 5
Quality of Coding 77.46%

77.46
45 GLM-5.2
Quality of Coding 77.12%

77.12
46 GPT-6 Luna
Quality of Coding 76.94%

76.94
47 GLM-4.5V
Quality of Coding 76.92%

76.92
48 Muse Spark 1.1
Quality of Coding 76.86%

76.86
49 Qwen2.5
Quality of Coding 76.78%

76.78
50 Qwen3.8-27B
Quality of Coding 76.23%

76.23
51 MiniMax M3.1 Flash
Quality of Coding 75.99%

75.99
52 Claude Opus 4.6
Quality of Coding 75.13%

75.13
53 Ring-flash-2.0
Quality of Coding 75.00%

75.00
54 Command A
Quality of Coding 74.85%

74.85
55 Qwen3.8-Flash-Next
Quality of Coding 74.66%

74.66
56 Muse Spark 1.2
Quality of Coding 74.66%

74.66
57 GPT-5.6 Luna
Quality of Coding 74.35%

74.35
58 GPT-5 nano
Quality of Coding 74.26%

74.26
59 Qwen3.7
Quality of Coding 74.16%

74.16
60 Qwen3.7-Plus
Quality of Coding 74.06%

74.06
61 Olmo 3.1 32B
Quality of Coding 73.67%

73.67
62 Kimi K2.6
Quality of Coding 73.34%

73.34
63 Mistral Large 4
Quality of Coding 72.76%

72.76
64 Muse Spark Open
Quality of Coding 72.64%

72.64
65 Muse Spark
Quality of Coding 72.44%

72.44
66 Gemini 3.5 Flash
Quality of Coding 71.91%

71.91
67 Gemini 3.6 Flash
Quality of Coding 71.82%

71.82
68 MiMo-V2.5-Pro
Quality of Coding 71.67%

71.67
69 Hy3
Quality of Coding 71.41%

71.41
70 Gemini 3.1 Pro
Quality of Coding 70.87%

70.87
71 Llama-3.3 Nemotron Super 49B v1
Quality of Coding 70.86%

70.86
72 Seed 2.0 Pro
Quality of Coding 70.63%

70.63
73 Olmo 3 32B Think
Quality of Coding 70.56%

70.56
74 Qwen3.6
Quality of Coding 69.79%

69.79
75 Claude Opus 4.5
Quality of Coding 69.63%

69.63
76 MiniMax M3
Quality of Coding 69.49%

69.49
77 Claude Sonnet 4.6
Quality of Coding 69.27%

69.27
78 GLM-5.1
Quality of Coding 69.24%

69.24
79 GPT-5.3 Codex
Quality of Coding 68.85%

68.85
80 GPT-5.4
Quality of Coding 68.64%

68.64
81 DeepSeek V3.1 Terminus
Quality of Coding 68.51%

68.51
82 Mistral Large
Quality of Coding 68.49%

68.49
83 Olmo 3.1 32B Think
Quality of Coding 67.75%

67.75
84 DeepSeek-V4-Flash
Quality of Coding 67.51%

67.51
85 Nemotron 3.5 Super VL
Quality of Coding 67.36%

67.36
86 GPT-5 mini
Quality of Coding 67.27%

67.27
87 MiMo-V2-Omni
Quality of Coding 66.72%

66.72
88 LongCat-Flash
Quality of Coding 66.60%

66.60
89 Qwen3.6 Plus
Quality of Coding 66.41%

66.41
90 Kimi K2
Quality of Coding 66.33%

66.33
91 Granite 4.0 H Small
Quality of Coding 66.12%

66.12
92 Llama 3.1 Nemotron 70B
Quality of Coding 65.83%

65.83
93 Step-3.5-Flash
Quality of Coding 65.47%

65.47
94 Claude Opus 4
Quality of Coding 65.17%

65.17
95 MiMo-V2.5
Quality of Coding 65.09%

65.09
96 Inkling-Small
Quality of Coding 65.08%

65.08
97 Gemini 3 Pro
Quality of Coding 64.29%

64.29
98 MiniMax M2.7
Quality of Coding 64.27%

64.27
99 GLM-5
Quality of Coding 64.16%

64.16
100 Granite 4.2 30B
Quality of Coding 64.15%

64.15