LLM leaderboard
Updated November 13, 2026

Point estimates with Supported and Estimated evidence labels.

#ModelEvidenceScore
1
Claude Fable 5.2
AnthropicEstimated
Estimated
93.65
2
Gemini 4 Carbon
GoogleSupported
Supported
89.15
3
Claude Opus 5.5
AnthropicSupported
Supported
86.62
4
GPT-6 Astra
OpenAISupported
Supported
85.65
5
GPT-6.1 Sol
OpenAISupported
Supported
84.6
6
Claude Sonnet 5.5
AnthropicSupported
Supported
83.05
7
Claude Fable 5.1
AnthropicSupported
Supported
79.21
8
Gemini 4 Argon
GoogleSupported
Supported
79.03
9
GPT-6.1 Terra
OpenAIEstimated
Estimated
79.64
10
GPT-6 Sol
OpenAISupported
Supported
77.41