Model IQ Index

Track coding-model benchmark scores and trends; score scales follow each source

All models

All upstream model scores and 30-day trends

Claude Sonnet 4 5 20250929
60
Volatile
Claude Opus 4.5
65
Good
Claude Opus 4 6
65
Good
Claude Sonnet 4.6
68
Good
Claude Opus 4.7
60
Volatile
Claude Opus 4.8
69
Good
Claude Fable 5
74
Good
Claude Sonnet 5
73
Good
Claude Opus 5
71
Good
DeepSeek V4 Flash
63
Volatile
DeepSeek V4 Pro
64
Volatile
GLM 5.2
57
Volatile
Gemini 3.1 Pro
75
Good
Gemini 3.1 Flash Lite
42
Volatile
Kimi 2.7
71
Good
Kimi K3
73
Good
GPT-5.3
71
Good
GPT 5.4
71
Good
GPT 5.5
71
Good
GPT-5.6 Sol
71
Good
GPT-5.6 Terra
67
Good
GPT-5.6 Luna
68
Good
anthropic
deepseek
glm
google
kimi
openai

Codex GPT-5.6 focus

Current IQ for each Codex configuration

Codex IQ data is temporarily unavailable

Claude Code focus

Current IQ for each Claude Code model configuration

Opus 4.8 xhigh

60

Opus 4.8 high

75

Opus 4.8 medium

60

Sonnet 5 max

120

Sonnet 5 xhigh

105

Sonnet 5 high

45

Fable 5 max

75

Fable 5 xhigh

120

Fable 5 high

90

Haiku 4.5

15