LLM Rankings: Top AI Models – Lewoniewski.info

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.

Last updated: October 19, 2026

Rank	Model Name	Relative Quality	Score
1	Claude Opus 5.5	Overall Quality 97.80%	97.80
Agents 100.00% Coding 99.35% Math 94.32% Reasoning 97.09%
2	GPT-6 Astra	Overall Quality 92.99%	92.99
Agents 93.90% Coding 93.48% Math 88.58% Reasoning 94.83%
3	Claude Sonnet 5.5	Overall Quality 92.58%	92.58
Agents 89.41% Coding 94.94% Math 91.48% Reasoning 93.05%
4	Gemini 4 Argon	Overall Quality 92.15%	92.15
Agents 84.40% Coding 89.78% Math 97.26% Reasoning 96.28%
5	Claude Fable 5.1	Overall Quality 92.14%	92.14
Agents 88.90% Coding 90.18% Math 96.87% Reasoning 93.11%
6	GPT-6.1 Sol	Overall Quality 89.40%	89.40
Agents 85.30% Coding 90.13% Math 89.96% Reasoning 91.05%
7	Claude Fable 5	Overall Quality 88.26%	88.26
Agents 73.45% Coding 89.38% Math 94.61% Reasoning 92.79%
8	Claude Opus 5	Overall Quality 88.19%	88.19
Agents 76.23% Coding 86.66% Math 95.07% Reasoning 93.10%
9	GPT-5.6 Sol	Overall Quality 84.99%	84.99
Agents 70.27% Coding 86.12% Math 88.53% Reasoning 91.30%
10	GPT-6 Sol	Overall Quality 84.59%	84.59
Agents 77.32% Coding 84.99% Math 90.25% Reasoning 85.25%
11	Muse Spark 1.3	Overall Quality 83.74%	83.74
Agents 63.26% Coding 86.25% Math 90.52% Reasoning 90.37%
12	Muse Wonder	Overall Quality 83.48%	83.48
Agents 68.54% Coding 86.31% Math 87.62% Reasoning 87.85%
13	DeepSeek-V4.1-Flash	Overall Quality 82.64%	82.64
Agents 69.51% Coding 87.47% Math 84.53% Reasoning 85.30%
14	Claude Opus 4.8	Overall Quality 82.32%	82.32
Agents 68.32% Coding 79.11% Math 89.98% Reasoning 89.76%
15	MiMo-V2.6-Pro	Overall Quality 82.02%	82.02
Agents 65.71% Coding 84.11% Math 93.57% Reasoning 83.10%
16	Kimi K3	Overall Quality 81.86%	81.86
Agents 62.79% Coding 80.62% Math 89.82% Reasoning 90.52%
17	Claude Opus 4.7	Overall Quality 81.40%	81.40
Agents 73.74% Coding 77.91% Math 88.33% Reasoning 85.37%
18	GLM-5.3	Overall Quality 80.81%	80.81
Agents 59.50% Coding 83.55% Math 88.79% Reasoning 86.95%
19	Muse Wonder Full	Overall Quality 80.56%	80.56
Agents 92.17% Coding 89.86% Math — Reasoning 83.88%
20	Gemini 3.8 Flash	Overall Quality 80.34%	80.34
Agents 57.91% Coding 79.88% Math 90.27% Reasoning 89.14%
21	GPT-5.5	Overall Quality 80.32%	80.32
Agents 61.24% Coding 78.20% Math 90.74% Reasoning 88.21%
22	Hy4	Overall Quality 80.15%	80.15
Agents 67.26% Coding 78.15% Math 86.03% Reasoning 86.83%
23	GPT-5.6 Terra	Overall Quality 79.62%	79.62
Agents 62.53% Coding 79.64% Math 85.43% Reasoning 87.13%
24	Claude Opus 4.6	Overall Quality 79.43%	79.43
Agents 68.70% Coding 75.13% Math 88.94% Reasoning 84.53%
25	Qwen3.8	Overall Quality 79.33%	79.33
Agents 61.50% Coding 78.47% Math 85.40% Reasoning 88.04%
26	DeepSeek-V4-Pro	Overall Quality 78.88%	78.88
Agents 64.24% Coding 78.30% Math 85.25% Reasoning 84.95%
27	Gemini 3.7 Flash	Overall Quality 78.43%	78.43
Agents 51.55% Coding 79.26% Math 89.73% Reasoning 87.98%
28	Grok 4.6	Overall Quality 78.39%	78.39
Agents 59.26% Coding 81.00% Math 81.26% Reasoning 86.63%
29	Muse Spark 1.1	Overall Quality 78.27%	78.27
Agents 53.21% Coding 76.86% Math 92.31% Reasoning 87.03%
30	Qwen3.8 Flash	Overall Quality 78.23%	78.23
Agents 60.87% Coding 80.71% Math 77.08% Reasoning 88.09%
31	Claude Sonnet 5	Overall Quality 77.67%	77.67
Agents 59.46% Coding 77.46% Math 84.27% Reasoning 85.61%
32	GLM-5.2	Overall Quality 77.49%	77.49
Agents 62.06% Coding 77.12% Math 86.11% Reasoning 82.41%
33	MiMo-V2.6-Flash	Overall Quality 77.46%	77.46
Agents 65.07% Coding 78.64% Math 90.60% Reasoning 75.78%
34	Gemini 3.5 Flash	Overall Quality 76.84%	76.84
Agents 70.43% Coding 71.91% Math 80.91% Reasoning 83.32%
35	Grok 4.7	Overall Quality 76.79%	76.79
Agents 64.90% Coding 81.99% Math 76.53% Reasoning 79.70%
36	Gemini 3.6 Flash	Overall Quality 76.39%	76.39
Agents 66.78% Coding 71.82% Math 84.80% Reasoning 81.77%
37	Grok 4.5	Overall Quality 76.35%	76.35
Agents 53.28% Coding 79.59% Math 81.06% Reasoning 85.36%
38	Kimi K2.6	Overall Quality 76.07%	76.07
Agents 67.13% Coding 73.34% Math 82.36% Reasoning 80.57%
39	GPT-6 Luna	Overall Quality 75.82%	75.82
Agents 61.63% Coding 76.94% Math 85.85% Reasoning 77.48%
40	Claude Haiku 5.5	Overall Quality 75.82%	75.82
Agents 65.65% Coding 79.53% Math 70.67% Reasoning 82.32%
41	GLM-5.3-Flash	Overall Quality 75.75%	75.75
Agents 52.87% Coding 77.94% Math 83.87% Reasoning 83.41%
42	Qwen3.7	Overall Quality 75.18%	75.18
Agents 50.36% Coding 74.16% Math 88.27% Reasoning 84.03%
43	Step 5	Overall Quality 75.08%	75.08
Agents 39.02% Coding 81.51% Math 89.03% Reasoning 83.40%
44	GPT-5.4	Overall Quality 74.98%	74.98
Agents 58.80% Coding 68.64% Math 87.42% Reasoning 83.79%
45	Qwen3.8-Flash-Next	Overall Quality 73.95%	73.95
Agents 58.19% Coding 74.66% Math 77.24% Reasoning 81.54%
46	Muse Spark 1.2	Overall Quality 73.88%	73.88
Agents 39.43% Coding 74.66% Math 92.95% Reasoning 83.37%
47	GPT-5.6 Luna	Overall Quality 73.65%	73.65
Agents 50.35% Coding 74.35% Math 82.40% Reasoning 82.64%
48	Gemini 3.1 Pro	Overall Quality 73.35%	73.35
Agents 43.25% Coding 70.87% Math 86.95% Reasoning 86.83%
49	Muse Spark	Overall Quality 73.26%	73.26
Agents 46.61% Coding 72.44% Math 89.10% Reasoning 81.30%
50	DeepSeek-V4-Flash	Overall Quality 73.05%	73.05
Agents 59.81% Coding 67.51% Math 81.67% Reasoning 81.66%
51	ChatGPT-4o	Overall Quality 72.94%	72.94
Agents — Coding 86.39% Math 80.41% Reasoning 69.81%
52	Qwen3.6	Overall Quality 72.85%	72.85
Agents — Coding 69.79% Math 91.69% Reasoning 78.57%
53	Qwen3.8-27B	Overall Quality 72.67%	72.67
Agents 47.75% Coding 76.23% Math 81.25% Reasoning 80.01%
54	ERNIE 5.0	Overall Quality 72.27%	72.27
Agents — Coding 89.64% Math 79.76% Reasoning 64.74%
55	GPT-5.5 Pro	Overall Quality 71.94%	71.94
Agents 74.96% Coding — Math 87.50% Reasoning 81.48%
56	Hy3	Overall Quality 71.90%	71.90
Agents 48.80% Coding 71.41% Math 86.35% Reasoning 78.17%
57	MiniMax M3.1 Flash	Overall Quality 71.69%	71.69
Agents 60.14% Coding 75.99% Math 63.89% Reasoning 80.30%
58	GLM-5.1	Overall Quality 71.24%	71.24
Agents 62.43% Coding 69.24% Math 75.62% Reasoning 76.19%
59	Seed 2.0 Pro	Overall Quality 71.23%	71.23
Agents 50.61% Coding 70.63% Math 81.88% Reasoning 78.49%
60	Qwen3.7-Plus	Overall Quality 71.13%	71.13
Agents 46.25% Coding 74.06% Math 78.56% Reasoning 79.83%
61	Gemini 3 Pro	Overall Quality 70.68%	70.68
Agents 47.65% Coding 64.29% Math 88.00% Reasoning 80.88%
62	Claude Sonnet 4.6	Overall Quality 70.67%	70.67
Agents 44.72% Coding 69.27% Math 84.78% Reasoning 79.95%
63	Muse Spark Open	Overall Quality 70.51%	70.51
Agents 48.25% Coding 72.64% Math 74.96% Reasoning 80.26%
64	Mistral Large 4	Overall Quality 70.26%	70.26
Agents 43.44% Coding 72.76% Math 81.02% Reasoning 78.48%
65	MiMo-V2.5	Overall Quality 69.87%	69.87
Agents 57.74% Coding 65.09% Math 85.74% Reasoning 72.17%
66	GLM-5	Overall Quality 69.76%	69.76
Agents 55.13% Coding 64.16% Math 86.68% Reasoning 73.82%
67	Kimi K2.5	Overall Quality 69.61%	69.61
Agents 52.35% Coding 63.68% Math 87.06% Reasoning 75.40%
68	Qwen3.6 Plus	Overall Quality 69.38%	69.38
Agents 56.70% Coding 66.41% Math 78.24% Reasoning 74.92%
69	MiMo-V2-Pro	Overall Quality 69.28%	69.28
Agents 50.26% Coding 63.98% Math 87.62% Reasoning 75.04%
70	Claude Opus 4.5	Overall Quality 69.20%	69.20
Agents 51.65% Coding 69.63% Math 74.35% Reasoning 77.02%
71	GPT-5.2	Overall Quality 69.05%	69.05
Agents 46.43% Coding 62.42% Math 83.91% Reasoning 80.85%
72	Gemini 3 Flash	Overall Quality 68.93%	68.93
Agents 51.65% Coding 62.54% Math 81.14% Reasoning 78.70%
73	GPT-5.5 Instant	Overall Quality 68.52%	68.52
Agents — Coding 84.93% Math 60.13% Reasoning 70.05%
74	Qwen3 VL 235B A22B	Overall Quality 68.06%	68.06
Agents 43.83% Coding 85.80% Math 76.62% Reasoning 60.78%
75	MiMo-V2.5-Pro	Overall Quality 67.63%	67.63
Agents 44.02% Coding 71.67% Math 78.89% Reasoning 71.81%
76	GPT-5.3 Codex	Overall Quality 67.43%	67.43
Agents 60.87% Coding 68.85% Math 58.71% Reasoning 76.19%
77	Kimi K2	Overall Quality 67.24%	67.24
Agents 48.52% Coding 66.33% Math 83.53% Reasoning 69.77%
78	MiniMax M2.7	Overall Quality 67.17%	67.17
Agents 48.35% Coding 64.27% Math 83.07% Reasoning 72.01%
79	LongCat-Flash	Overall Quality 66.95%	66.95
Agents 43.48% Coding 66.60% Math 79.21% Reasoning 74.78%
80	Qwen3.5-397B-A17B	Overall Quality 66.95%	66.95
Agents 55.83% Coding 60.96% Math 74.99% Reasoning 75.00%
81	MiniMax M3	Overall Quality 66.60%	66.60
Agents 34.49% Coding 69.49% Math 74.74% Reasoning 79.68%
82	GPT-5.1	Overall Quality 66.40%	66.40
Agents — Coding 59.46% Math 81.96% Reasoning 73.89%
83	INTELLECT-3	Overall Quality 66.37%	66.37
Agents — Coding 78.70% Math 75.55% Reasoning 58.83%
84	Muse Glimmer	Overall Quality 66.24%	66.24
Agents — Coding 58.86% Math 87.15% Reasoning 70.50%
85	DeepSeek V3.1 Terminus	Overall Quality 65.93%	65.93
Agents — Coding 68.51% Math 79.31% Reasoning 65.06%
86	Step-3.5-Flash	Overall Quality 65.85%	65.85
Agents 47.83% Coding 65.47% Math 79.84% Reasoning 68.93%
87	Grok-4	Overall Quality 65.28%	65.28
Agents — Coding 61.48% Math 79.29% Reasoning 69.93%
88	GPT-5.4 mini	Overall Quality 65.03%	65.03
Agents 54.43% Coding 62.50% Math 68.96% Reasoning 72.01%
89	GPT-5.2 Pro	Overall Quality 64.86%	64.86
Agents 52.35% Coding — Math 83.78% Reasoning 75.45%
90	MiniMax M2.5	Overall Quality 64.64%	64.64
Agents 51.30% Coding 60.42% Math 78.68% Reasoning 68.38%
91	Kimi K2.7 Code	Overall Quality 64.42%	64.42
Agents 66.43% Coding 62.77% Math 55.94% Reasoning 70.38%
92	MiMo-V2-Omni	Overall Quality 64.37%	64.37
Agents 35.83% Coding 66.72% Math 85.11% Reasoning 67.23%
93	Mistral Medium 3	Overall Quality 64.34%	64.34
Agents — Coding 81.21% Math 71.63% Reasoning 52.15%
94	Qwen3.5-122B-A10B	Overall Quality 63.99%	63.99
Agents 48.17% Coding 55.59% Math 82.98% Reasoning 70.26%
95	GLM-4.5V	Overall Quality 63.76%	63.76
Agents — Coding 76.92% Math 73.67% Reasoning 53.18%
96	Inkling-Small	Overall Quality 63.59%	63.59
Agents 32.89% Coding 65.08% Math 76.33% Reasoning 74.08%
97	GLM-4.7	Overall Quality 63.10%	63.10
Agents 37.39% Coding 58.65% Math 82.04% Reasoning 72.06%
98	Qwen2.5	Overall Quality 62.92%	62.92
Agents — Coding 76.78% Math 73.82% Reasoning 50.41%
99	GPT-5	Overall Quality 62.82%	62.82
Agents 36.52% Coding 62.71% Math 75.88% Reasoning 71.75%
100	Ling 3.0 Flash	Overall Quality 62.75%	62.75
Agents 54.43% Coding 59.84% Math 72.37% Reasoning 64.79%