llm.ing

AA Intelligence Index

run by Artificial Analysisunit: indexDisplay with attribution; no redistribution95 modelssource ↗methodology ↗

Composite of independently run evals by Artificial Analysis.

AA Intelligence IndexAA Coding IndexAA Agentic IndexMedian output speedMedian time to first token
#ModelScoreObserved
1Claude Fable 5
59.9
mirror
Jul 23, 2026
2GPT-5.6 Sol
58.9
mirror
Jul 23, 2026
3Kimi K3
57.1
mirror
Jul 23, 2026
4Claude Opus 4.8
55.7
mirror
Jul 23, 2026
5GPT-5.6 Terra
55.0
mirror
Jul 23, 2026
6GPT-5.5
54.8
independent
Jul 24, 2026
7Grok 4.5
53.8
mirror
Jul 23, 2026
8Claude Opus 4.7
53.5
mirror
Jul 23, 2026
9Claude Sonnet 5
53.4
independent
Jul 24, 2026
10GPT-5.4
51.4
mirror
Jul 23, 2026
11GPT-5.6 Luna
51.2
independent
Jul 24, 2026
12GLM-5.2
51.1
mirror
Jul 23, 2026
13Muse Spark 1.1
50.6
mirror
Jul 23, 2026
14Gemini 3.5 Flash
50.2
independent
Jul 24, 2026
15Gemini 3.6 Flash
50.1
mirror
Jul 23, 2026
16Claude Sonnet 4.6
47.2
mirror
Jul 23, 2026
17Gemini 3.1 Pro Preview
46.5
mirror
Jul 23, 2026
18Qwen3.7 Max
46.0
mirror
Jul 23, 2026
19MiniMax-M3
44.4
independent
Jul 24, 2026
20DeepSeek V4 Pro
44.3
mirror
Jul 23, 2026
21Kimi K2.6
44.2
mirror
Jul 23, 2026
22GPT-5.2
42.2
independent
Jul 24, 2026
23Kimi K2.7 Code
41.9
mirror
Jul 23, 2026
24Claude Opus 4.5 (latest)thinking
40.8
independent
Jul 24, 2026
25DeepSeek V4 Flash
40.3
mirror
Jul 23, 2026
26GLM-5.1
40.2
mirror
Jul 23, 2026
27GPT-5.2 Codex
40.1
independent
Jul 24, 2026
28GPT-5.4 mini
40.0
mirror
Jul 23, 2026
29Grok Build 0.1
39.8
independent
Jul 24, 2026
30Qwen3.6 Plus
39.6
mirror
Jul 23, 2026
31GLM-5
39.5
independent
Jul 24, 2026
32Qwen3.7 Plus
39.0
mirror
Jul 23, 2026
33GPT-5.4 nano
38.2
mirror
Jul 23, 2026
34MiniMax-M2.7
38.1
independent
Jul 24, 2026
35Claude Opus 4.6
37.8
independent
Jul 24, 2026
35Nemotron 3 Ultra 550B A55B
37.8
mirror
Jul 23, 2026
37Grok 4.3
37.6
mirror
Jul 23, 2026
38Qwen3.6 27B
37.1
mirror
Jul 23, 2026
39GPT-5.1
36.9
independent
Jul 24, 2026
40Gemini 3.5 Flash Lite
36.5
independent
Jul 24, 2026
41Claude Sonnet 4.5 (latest)
36.4
mirror
Jul 23, 2026
42Kimi K2.5
35.4
mirror
Jul 23, 2026
43GPT-5
34.7
independent
Jul 24, 2026
43GPT-5.1 Codex
34.7
independent
Jul 24, 2026
45GLM-5V-Turbo
34.5
independent
Jul 24, 2026
46Qwen3.5 27B
33.8
independent
Jul 24, 2026
47Qwen3.5 397B-A17B
33.7
independent
Jul 24, 2026
47GLM-4.7
33.7
mirror
Jul 23, 2026
47MiniMax-M2.5
33.7
independent
Jul 24, 2026
50Qwen3.5 122B-A10B
32.3
independent
Jul 24, 2026
51DeepSeek V3.2
32.0
mirror
Jul 23, 2026
52Qwen3.6 35B-A3B
31.6
mirror
Jul 23, 2026
53MiniMax-M2.1
31.4
independent
Jul 24, 2026
54o3
30.4
independent
Jul 24, 2026
55Step 3.7 Flash
30.3
mirror
Jul 23, 2026
56Claude Haiku 4.5 (latest)
29.6
mirror
Jul 23, 2026
57Gemma 4 31B IT
29.4
mirror
Jul 23, 2026
58Qwen3.5 35B-A3B
29.3
independent
Jul 24, 2026
59GLM-4.6
28.7
mirror
Jul 23, 2026
60Gemini 2.5 Pro
25.8
mirror
Jul 23, 2026
61Gemma 4 26B A4B IT
25.7
mirror
Jul 23, 2026
62Nemotron 3 Super
25.4
mirror
Jul 23, 2026
63GPT-5 Mini
25.3
independent
Jul 24, 2026
64Gemini 3.1 Flash Lite Preview
25.0
mirror
Jul 23, 2026
65GLM-4.7-Flash
22.9
independent
Jul 24, 2026
66Mercury 2
21.4
independent
Jul 24, 2026
67DeepSeek V3.1thinking
20.7
independent
Jul 24, 2026
68Gemini 2.5 Flashthinking
20.1
independent
Jul 24, 2026
69Qwen3 235B-A22Bthinking
19.6
mirror
Jul 23, 2026
70GLM-4.5
19.5
independent
Jul 24, 2026
71GPT-4.1
19.4
independent
Jul 24, 2026
72Devstral 2
19.2
mirror
Jul 23, 2026
73o3-mini
19.0
independent
Jul 24, 2026
74DeepSeek R1
18.5
independent
Jul 24, 2026
75Qwen3-Coder 480B-A35B Instruct
18.0
independent
Jul 24, 2026
76Sonar Reasoning Pro
17.8
independent
Jul 24, 2026
77Kimi K2 Thinkingthinking
17.3
mirror
Jul 23, 2026
78Qwen3-Next 80B-A3B (Thinking)thinking
16.7
mirror
Jul 23, 2026
79GPT-4.1 mini
14.8
mirror
Jul 23, 2026
80nemotron-3-nano-30b-a3b
14.2
mirror
Jul 23, 2026
81Qwen3-Next 80B-A3B Instruct
13.7
independent
Jul 24, 2026
82Sonarthinking
11.7
independent
Jul 24, 2026
83Qwen3 32B
11.5
mirror
Jul 23, 2026
84GLM-4.6V
11.0
independent
Jul 24, 2026
85Magistral Small
10.7
independent
Jul 24, 2026
86Qwen3 14B
10.4
mirror
Jul 23, 2026
87GPT-4.1 nano
9.6
mirror
Jul 23, 2026
88Llama-3.3-70B-Instruct
9.4
mirror
Jul 23, 2026
89Sonar Pro
9.3
independent
Jul 24, 2026
90GLM-4.5Vthinking
9.1
independent
Jul 24, 2026
91Qwen3 8B
8.3
mirror
Jul 23, 2026
92Qwen2.5 32B Instruct
7.5
independent
Jul 24, 2026
93nvidia-nemotron-nano-9b-v2
7.4
independent
Jul 24, 2026
94Gemini 2.5 Flash-Lite
6.9
independent
Jul 24, 2026
95Qwen Turbo
6.3
independent
Jul 24, 2026

Bars on a 0–60 index scale. Scores link to model pages; every number keeps its source on the model page.