The Agents Leaderboard ranks large language models by their ability to act as autonomous problem-solvers rather than simply generate responses. Unlike standard text-generation evaluations, agentic benchmarks test whether a model can independently break down complex goals, execute shell commands in terminal environments, invoke external APIs, interpret tool outputs, and recover from runtime errors.

Drawing on evaluations agent-focused benchmarks this leaderboard highlights models capable of handling complex, multi-step tasks across developer and enterprise workflows.

Looking for well-rounded performance? Agentic capabilities account for 20% of the composite evaluation. Check the Overall LLM Ranking to see how top agentic models perform across coding, mathematics, and complex reasoning.

Last updated: October 9, 2026
Rank Model Name Relative Quality Score
1 Claude Opus 5.5
Quality of Agents 97.74%

97.74
2 GPT-6 Astra
Quality of Agents 90.21%

90.21
3 Claude Fable 5.1
Quality of Agents 89.71%

89.71
4 Gemini 4 Argon
Quality of Agents 87.36%

87.36
5 Ling 3.1 Flash
Quality of Agents 87.26%

87.26
6 GPT-6.1 Sol
Quality of Agents 84.00%

84.00
7 Claude Fable 5
Quality of Agents 83.95%

83.95
8 Claude Opus 5
Quality of Agents 83.91%

83.91
9 DeepSeek-V4.1-Flash
Quality of Agents 83.90%

83.90
10 Claude Sonnet 5.5
Quality of Agents 80.26%

80.26
11 DeepSeek-V4-Flash
Quality of Agents 79.37%

79.37
12 GPT-6 Sol
Quality of Agents 79.29%

79.29
13 GPT-5.6 Sol
Quality of Agents 78.50%

78.50
14 Muse Spark 1.3
Quality of Agents 78.41%

78.41
15 Kimi K3
Quality of Agents 76.95%

76.95
16 Qwen3.8
Quality of Agents 75.55%

75.55
17 Seed 2.1 Pro
Quality of Agents 74.87%

74.87
18 GLM-5.3
Quality of Agents 74.42%

74.42
19 MiMo-V2.6-Pro
Quality of Agents 74.42%

74.42
20 Claude Opus 4.8
Quality of Agents 74.22%

74.22
21 Claude Sonnet 5
Quality of Agents 73.08%

73.08
22 Hy4
Quality of Agents 73.07%

73.07
23 Claude Haiku 5.5
Quality of Agents 72.78%

72.78
24 GPT-5.5
Quality of Agents 70.97%

70.97
25 Grok 4.7
Quality of Agents 70.92%

70.92
26 Ling 3.0 Flash Fin
Quality of Agents 70.51%

70.51
27 Gemini 3.8 Flash
Quality of Agents 70.03%

70.03
28 DeepSeek-V4-Pro
Quality of Agents 69.99%

69.99
29 Grok 4.6
Quality of Agents 69.94%

69.94
30 GPT-5.6 Terra
Quality of Agents 69.76%

69.76
31 Claude Opus 4.7
Quality of Agents 69.44%

69.44
32 Qwen3.8-Flash-Next
Quality of Agents 69.03%

69.03
33 GLM-5.3-Flash
Quality of Agents 67.82%

67.82
34 GLM-5.2
Quality of Agents 67.24%

67.24
35 Gemini 3.5 Flash
Quality of Agents 66.86%

66.86
36 Grok 4.5
Quality of Agents 66.75%

66.75
37 GPT-6 Luna
Quality of Agents 66.42%

66.42
38 Claude Opus 4.6
Quality of Agents 65.73%

65.73
39 Qwen3.8-27B
Quality of Agents 65.44%

65.44
40 MiMo-V2.6-Flash
Quality of Agents 65.25%

65.25
41 Gemini 3.7 Flash
Quality of Agents 64.35%

64.35
42 GPT-5.4
Quality of Agents 64.18%

64.18
43 Kimi K2.6
Quality of Agents 63.67%

63.67
44 Muse Spark 1.1
Quality of Agents 62.54%

62.54
45 Kimi K2.7 Code
Quality of Agents 62.37%

62.37
46 GLM-5.1
Quality of Agents 61.95%

61.95
47 GPT-5.6 Luna
Quality of Agents 61.09%

61.09
48 Mistral Large 4
Quality of Agents 60.86%

60.86
49 GPT-5.3 Codex
Quality of Agents 59.86%

59.86
50 Grok Build 0.1
Quality of Agents 59.25%

59.25
51 Muse Spark 1.2
Quality of Agents 59.10%

59.10
52 GPT-5.2 Codex
Quality of Agents 58.74%

58.74
53 Claude Sonnet 4.6
Quality of Agents 58.53%

58.53
54 MiMo-V2.5
Quality of Agents 57.24%

57.24
55 GPT-5.2
Quality of Agents 55.57%

55.57
56 Qwen3.5-397B-A17B
Quality of Agents 55.32%

55.32
57 Qwen3.6 Plus
Quality of Agents 54.79%

54.79
58 GLM-5
Quality of Agents 54.62%

54.62
59 Step 5
Quality of Agents 54.26%

54.26
60 Ling 3.0 Flash
Quality of Agents 53.93%

53.93
61 GPT-5.4 mini
Quality of Agents 53.76%

53.76
62 GLM-5V-Turbo
Quality of Agents 53.75%

53.75
63 Qwen3.7
Quality of Agents 53.41%

53.41
64 Gemini 3.5 Flash-Lite
Quality of Agents 52.69%

52.69
65 MiniMax M2.1
Quality of Agents 52.01%

52.01
66 Kimi K2.5
Quality of Agents 51.83%

51.83
67 Claude Opus 4.5
Quality of Agents 51.25%

51.25
68 Gemini 3 Flash
Quality of Agents 51.13%

51.13
69 Hy3
Quality of Agents 51.07%

51.07
70 Claude Sonnet 4.5
Quality of Agents 50.61%

50.61
71 MiniMax M2.5
Quality of Agents 50.44%

50.44
72 Seed 2.0 Pro
Quality of Agents 50.26%

50.26
73 MiMo-V2-Pro
Quality of Agents 50.26%

50.26
74 GPT-5.4 nano
Quality of Agents 49.99%

49.99
75 Qwen3.6-27B
Quality of Agents 49.59%

49.59
76 Gemini 3.6 Flash
Quality of Agents 49.39%

49.39
77 MiniMax M3
Quality of Agents 48.81%

48.81
78 Kimi K2
Quality of Agents 48.69%

48.69
79 Gemini 3.1 Pro
Quality of Agents 48.31%

48.31
80 MiniMax M2.7
Quality of Agents 47.82%

47.82
81 Step-3.5-Flash
Quality of Agents 47.47%

47.47
82 Qwen3.5-122B-A10B
Quality of Agents 47.47%

47.47
83 Gemini 3 Pro
Quality of Agents 47.29%

47.29
84 Qwen3.5-27B
Quality of Agents 45.90%

45.90
85 Qwen3.7-Plus
Quality of Agents 45.34%

45.34
86 Muse Spark
Quality of Agents 45.20%

45.20
87 Gemma 4 31B
Quality of Agents 44.33%

44.33
88 MiMo-V2.5-Pro
Quality of Agents 44.19%

44.19
89 Claude Haiku 4.5
Quality of Agents 43.98%

43.98
90 Inkling
Quality of Agents 43.70%

43.70
91 Qwen3 VL 235B A22B
Quality of Agents 43.46%

43.46
92 LongCat-Flash
Quality of Agents 43.46%

43.46
93 Claude Opus 4.1
Quality of Agents 43.46%

43.46
94 Step 3.7 Flash
Quality of Agents 43.46%

43.46
95 Qwen3 VL 32B
Quality of Agents 43.11%

43.11
96 Gemma 4 26B-A4B
Quality of Agents 42.76%

42.76
97 Qwen3.5-35B-A3B
Quality of Agents 42.23%

42.23
98 Qwen3-Next-80B-A3B
Quality of Agents 42.06%

42.06
99 DeepSeek-V3.2
Quality of Agents 41.71%

41.71
100 Qwen3.6-35B-A3B
Quality of Agents 41.01%

41.01