LLM Rankings: Top AI Models

This leaderboard presents the top 100 Large Language Models (LLMs) ranked by their Relative Quality Score. Unlike single-source lists, this ranking utilizes a unified, synthetic aggregation model to deliver a comprehensive and unbiased picture of current AI capabilities.

The ranking includes only those AI models that have been evaluated by at least two widely recognized independent benchmarking sources. Each model’s overall score is calculated as a composite across four key domains:

🤖 Agents (20%) — autonomous execution, environment interaction, and tool usage;
💻 Coding (30%) — measuring programming ability, syntax accuracy, and algorithmic logic;
🧮 Math (20%) — testing quantitative problem-solving;
🔬 Reasoning (30%) — assessing logical thinking and multi-step deduction.

Clicking on any model expands details across all four sub-aspects.

Rank	Model Name	Relative Quality	Score
1	Claude Opus 5.5	Overall Quality 97.73%	97.73
Agents 97.90% Coding 99.24% Math 94.90% Reasoning 97.99%
2	Gemini 4 Carbon	Overall Quality 96.91%	96.91
Agents 95.55% Coding 96.23% Math 98.91% Reasoning 97.17%
3	GPT-6 Astra	Overall Quality 93.19%	93.19
Agents 94.17% Coding 92.86% Math 88.05% Reasoning 96.29%
4	Claude Sonnet 5.5	Overall Quality 93.09%	93.09
Agents 88.73% Coding 94.70% Math 91.51% Reasoning 95.44%
5	Claude Fable 5.1	Overall Quality 92.43%	92.43
Agents 90.37% Coding 90.21% Math 97.00% Reasoning 92.97%
6	Gemini 4 Argon	Overall Quality 91.36%	91.36
Agents 82.30% Coding 89.28% Math 97.26% Reasoning 95.55%
7	GPT-6.1 Sol	Overall Quality 90.27%	90.27
Agents 85.37% Coding 89.86% Math 90.55% Reasoning 93.74%
8	Claude Fable 5	Overall Quality 88.23%	88.23
Agents 78.45% Coding 86.20% Math 94.32% Reasoning 92.74%
9	Claude Opus 5	Overall Quality 88.10%	88.10
Agents 77.85% Coding 85.08% Math 94.96% Reasoning 93.38%
10	GPT-6.1 Terra	Overall Quality 86.88%	86.88
Agents 76.63% Coding 84.05% Math 97.01% Reasoning 89.80%
11	GPT-5.6 Sol	Overall Quality 85.32%	85.32
Agents 74.30% Coding 83.84% Math 88.13% Reasoning 92.27%
12	Kimi K3.1 Pro	Overall Quality 85.27%	85.27
Agents 68.86% Coding 86.15% Math 85.92% Reasoning 94.91%
13	Muse Wonder	Overall Quality 84.88%	84.88
Agents 70.11% Coding 86.20% Math 91.40% Reasoning 89.06%
14	Claude Opus 4.8	Overall Quality 84.68%	84.68
Agents 79.45% Coding 78.46% Math 89.82% Reasoning 90.96%
15	Muse Spark 1.3	Overall Quality 83.44%	83.44
Agents 58.99% Coding 86.15% Math 90.64% Reasoning 92.22%
16	GLM-5.3	Overall Quality 83.43%	83.43
Agents 68.30% Coding 82.86% Math 89.12% Reasoning 90.29%
17	GPT-6 Sol	Overall Quality 83.29%	83.29
Agents 74.22% Coding 82.03% Math 89.88% Reasoning 86.21%
18	DeepSeek-V4.1-Flash	Overall Quality 83.02%	83.02
Agents 63.54% Coding 87.02% Math 85.38% Reasoning 90.43%
19	MiMo-V2.6-Pro	Overall Quality 82.36%	82.36
Agents 62.18% Coding 83.62% Math 93.45% Reasoning 87.17%
20	GPT-5.5	Overall Quality 82.13%	82.13
Agents 75.17% Coding 75.38% Math 90.51% Reasoning 87.94%
21	Claude Opus 4.7	Overall Quality 82.03%	82.03
Agents 72.95% Coding 76.87% Math 88.17% Reasoning 89.15%
22	Kimi K3	Overall Quality 81.44%	81.44
Agents 57.54% Coding 80.28% Math 89.85% Reasoning 92.93%
23	Hy4	Overall Quality 80.81%	80.81
Agents 72.10% Coding 77.35% Math 85.60% Reasoning 86.88%
24	Claude Opus 4.6	Overall Quality 80.24%	80.24
Agents 68.32% Coding 74.03% Math 88.64% Reasoning 88.80%
25	Gemini 3.8 Flash	Overall Quality 79.56%	79.56
Agents 53.46% Coding 79.57% Math 90.17% Reasoning 89.87%
26	DeepSeek-V4-Pro	Overall Quality 79.48%	79.48
Agents 64.93% Coding 77.59% Math 84.98% Reasoning 87.41%
27	MiMo-V2.6-Flash	Overall Quality 79.38%	79.38
Agents 63.40% Coding 78.10% Math 89.24% Reasoning 84.73%
28	Gemini 3.7 Flash	Overall Quality 79.35%	79.35
Agents 60.67% Coding 78.15% Math 89.12% Reasoning 86.49%
29	Qwen3.8	Overall Quality 79.33%	79.33
Agents 57.87% Coding 78.23% Math 85.00% Reasoning 90.95%
30	GPT-5.6 Terra	Overall Quality 79.18%	79.18
Agents 57.22% Coding 79.04% Math 85.10% Reasoning 90.03%
31	Muse Spark 1.1	Overall Quality 79.14%	79.14
Agents 53.65% Coding 76.25% Math 92.18% Reasoning 90.33%
32	Claude Sonnet 5	Overall Quality 78.69%	78.69
Agents 64.53% Coding 75.21% Math 84.31% Reasoning 87.87%
33	Grok 4.5	Overall Quality 78.66%	78.66
Agents 72.95% Coding 74.90% Math 80.78% Reasoning 84.81%
34	Grok 4.6	Overall Quality 78.45%	78.45
Agents 56.12% Coding 80.82% Math 80.89% Reasoning 89.35%
35	Claude Haiku 5.5	Overall Quality 78.36%	78.36
Agents 60.92% Coding 81.94% Math 79.08% Reasoning 85.92%
36	GLM-5.2	Overall Quality 78.25%	78.25
Agents 61.41% Coding 76.56% Math 86.01% Reasoning 85.99%
37	GPT-5.6 Luna	Overall Quality 78.19%	78.19
Agents 72.77% Coding 73.59% Math 82.07% Reasoning 83.81%
38	Muse Wonder Full	Overall Quality 77.99%	77.99
Agents 82.79% Coding 89.81% Math — Reasoning 81.63%
39	Grok 4.7	Overall Quality 77.95%	77.95
Agents 65.16% Coding 81.60% Math 76.26% Reasoning 83.95%
40	Kimi K2.6	Overall Quality 77.30%	77.30
Agents 67.47% Coding 72.50% Math 82.02% Reasoning 85.51%
41	DeepSeek-V4-Flash	Overall Quality 77.24%	77.24
Agents 72.60% Coding 74.09% Math 81.40% Reasoning 80.71%
42	Gemini 3.5 Flash	Overall Quality 76.95%	76.95
Agents 70.38% Coding 71.21% Math 80.77% Reasoning 84.52%
43	Gemini 3.6 Flash	Overall Quality 76.51%	76.51
Agents 66.78% Coding 71.23% Math 84.60% Reasoning 82.89%
44	GPT-6 Luna	Overall Quality 76.47%	76.47
Agents 56.59% Coding 76.90% Math 85.89% Reasoning 83.02%
45	GPT-5.4	Overall Quality 76.44%	76.44
Agents 66.27% Coding 67.78% Math 87.21% Reasoning 84.70%
46	Step 5	Overall Quality 76.32%	76.32
Agents 35.93% Coding 81.31% Math 89.70% Reasoning 89.32%
47	Mistral Large 4	Overall Quality 76.20%	76.20
Agents 51.06% Coding 84.96% Math 74.89% Reasoning 85.09%
48	GLM-5.3-Flash	Overall Quality 76.10%	76.10
Agents 50.23% Coding 77.65% Math 83.72% Reasoning 86.74%
49	Beam 1.0	Overall Quality 76.09%	76.09
Agents 70.21% Coding 71.71% Math 80.44% Reasoning 81.49%
50	Qwen3.7	Overall Quality 75.98%	75.98
Agents 51.47% Coding 73.03% Math 87.66% Reasoning 87.47%
51	ERNIE 5.0	Overall Quality 75.13%	75.13
Agents — Coding 87.97% Math 79.52% Reasoning 76.10%
52	Qwen3.8-Flash-Next	Overall Quality 74.84%	74.84
Agents 58.00% Coding 74.10% Math 76.37% Reasoning 85.77%
53	Gemini 3.1 Pro	Overall Quality 74.83%	74.83
Agents 50.99% Coding 70.91% Math 86.85% Reasoning 86.63%
54	Claude Sonnet 4.6	Overall Quality 74.66%	74.66
Agents 61.99% Coding 68.92% Math 84.45% Reasoning 82.31%
55	Muse Spark 1.2	Overall Quality 74.12%	74.12
Agents 47.21% Coding 73.22% Math 93.76% Reasoning 79.87%
56	Muse Spark	Overall Quality 73.90%	73.90
Agents 47.60% Coding 70.99% Math 89.06% Reasoning 84.24%
57	Qwen3.8-27B	Overall Quality 73.30%	73.30
Agents 46.40% Coding 75.57% Math 80.38% Reasoning 84.23%
58	Hy3	Overall Quality 73.25%	73.25
Agents 48.74% Coding 70.32% Math 86.40% Reasoning 83.74%
59	GLM-5.1	Overall Quality 72.63%	72.63
Agents 62.33% Coding 68.47% Math 75.35% Reasoning 81.84%
60	MiniMax M3.1 Flash	Overall Quality 72.26%	72.26
Agents 54.51% Coding 78.01% Math 63.54% Reasoning 84.17%
61	Qwen3.7-Plus	Overall Quality 72.03%	72.03
Agents 46.83% Coding 72.93% Math 78.22% Reasoning 83.80%
62	GPT-5.5 Pro	Overall Quality 71.90%	71.90
Agents 73.80% Coding — Math 87.64% Reasoning 82.04%
63	ChatGPT-4o	Overall Quality 71.83%	71.83
Agents — Coding 84.64% Math 79.88% Reasoning 68.22%
64	Seed 2.0 Pro	Overall Quality 71.68%	71.68
Agents 50.51% Coding 69.22% Math 81.32% Reasoning 81.84%
65	DeepSeek V3.1 Terminus	Overall Quality 71.17%	71.17
Agents — Coding 67.64% Math 78.94% Reasoning 83.62%
66	MiMo-V2.5	Overall Quality 71.09%	71.09
Agents 56.85% Coding 63.93% Math 85.34% Reasoning 78.24%
67	GLM-5	Overall Quality 71.01%	71.01
Agents 54.45% Coding 63.23% Math 86.27% Reasoning 79.65%
68	Kimi K2.5	Overall Quality 70.94%	70.94
Agents 52.57% Coding 62.78% Math 86.63% Reasoning 80.90%
69	Qwen3.6 Plus	Overall Quality 70.90%	70.90
Agents 56.34% Coding 65.34% Math 77.78% Reasoning 81.57%
70	Claude Opus 4.5	Overall Quality 70.75%	70.75
Agents 51.20% Coding 68.49% Math 74.15% Reasoning 83.76%
71	Muse Glimmer	Overall Quality 70.60%	70.60
Agents — Coding 58.26% Math 86.27% Reasoning 86.23%
72	Gemini 3 Pro	Overall Quality 70.50%	70.50
Agents 47.09% Coding 63.31% Math 87.65% Reasoning 81.85%
73	MiMo-V2-Pro	Overall Quality 70.46%	70.46
Agents 50.34% Coding 62.71% Math 87.36% Reasoning 80.36%
74	GPT-5.5 Instant	Overall Quality 70.15%	70.15
Agents — Coding 84.06% Math 59.96% Reasoning 76.46%
75	Qwen3 VL 235B A22B	Overall Quality 70.10%	70.10
Agents 43.49% Coding 84.06% Math 76.24% Reasoning 69.79%
76	Muse Spark Open	Overall Quality 69.90%	69.90
Agents 44.75% Coding 72.52% Math 76.27% Reasoning 79.79%
77	GPT-5.2	Overall Quality 69.24%	69.24
Agents 45.89% Coding 61.35% Math 83.63% Reasoning 83.09%
78	Mistral Medium 3.1	Overall Quality 69.15%	69.15
Agents — Coding 62.14% Math 78.63% Reasoning 82.61%
79	Gemini 3 Flash	Overall Quality 69.03%	69.03
Agents 51.37% Coding 61.56% Math 80.92% Reasoning 80.36%
80	MiniMax M3	Overall Quality 68.86%	68.86
Agents 44.64% Coding 68.02% Math 74.36% Reasoning 82.19%
81	Kimi K2	Overall Quality 68.75%	68.75
Agents 47.43% Coding 65.38% Math 83.28% Reasoning 76.65%
82	MiniMax M2.7	Overall Quality 68.18%	68.18
Agents 48.12% Coding 63.43% Math 82.84% Reasoning 76.55%
83	Qwen3.5-397B-A17B	Overall Quality 67.97%	67.97
Agents 55.14% Coding 60.20% Math 74.66% Reasoning 79.83%
84	LongCat-Flash	Overall Quality 67.61%	67.61
Agents 43.32% Coding 65.46% Math 78.90% Reasoning 78.42%
85	GPT-5.1	Overall Quality 67.46%	67.46
Agents — Coding 58.54% Math 81.71% Reasoning 78.52%
86	MiMo-V2.5-Pro	Overall Quality 67.37%	67.37
Agents 44.63% Coding 70.42% Math 78.49% Reasoning 72.06%
87	Step-3.5-Flash	Overall Quality 67.33%	67.33
Agents 46.92% Coding 64.25% Math 79.61% Reasoning 75.82%
88	MiniMax M2.5	Overall Quality 66.32%	66.32
Agents 50.68% Coding 59.48% Math 78.32% Reasoning 75.60%
89	GPT-5.4 mini	Overall Quality 65.70%	65.70
Agents 54.11% Coding 61.75% Math 68.41% Reasoning 75.56%
90	Ling 3.0 Flash Fin	Overall Quality 65.66%	65.66
Agents 70.21% Coding 58.33% Math — Reasoning 80.39%
91	Grok-4	Overall Quality 65.51%	65.51
Agents — Coding 60.34% Math 79.06% Reasoning 71.98%
92	Qwen3.5-122B-A10B	Overall Quality 65.12%	65.12
Agents 47.60% Coding 54.86% Math 82.65% Reasoning 75.37%
93	GPT-5.2 Pro	Overall Quality 65.02%	65.02
Agents 52.23% Coding — Math 83.71% Reasoning 76.10%
94	K2 Horizon 375B-A23B	Overall Quality 64.91%	64.91
Agents 61.82% Coding 56.47% Math 72.14% Reasoning 70.58%
95	MiMo-V2-Omni	Overall Quality 64.87%	64.87
Agents 34.76% Coding 64.85% Math 84.71% Reasoning 71.74%
96	Claude Sonnet 4.5	Overall Quality 64.41%	64.41
Agents 50.34% Coding 60.11% Math 66.21% Reasoning 76.89%
97	Inkling-Small	Overall Quality 64.40%	64.40
Agents 33.28% Coding 64.30% Math 75.44% Reasoning 77.90%
98	Trinity Large	Overall Quality 64.06%	64.06
Agents — Coding 47.86% Math 76.76% Reasoning 81.16%
99	GLM-4.7	Overall Quality 63.86%	63.86
Agents 36.30% Coding 57.84% Math 81.72% Reasoning 76.35%
100	Mistral Small 3.2	Overall Quality 63.85%	63.85
Agents — Coding 56.59% Math 69.58% Reasoning 76.52%