BenchLM Leaderboard
BenchAlign leaderboard with Supported and Estimated evidence labels
#ModelCreatorEvidenceScore
1Claude Opus 5.5AnthropicSupported86.56
2GPT-6.1 SolOpenAISupported84.07
3GPT-6 AstraOpenAISupported83.88
4Claude Fable 5.1AnthropicSupported83.18
5Claude Sonnet 5.5AnthropicSupported82.18
6Claude Fable 5AnthropicSupported81.27
7Gemini 4 ArgonGoogleSupported80.97
8Claude Opus 5AnthropicSupported79.63
9GPT-6 SolOpenAISupported78.48
10GPT-5.6 SolOpenAISupported75.32
LLM benchmark leaderboard by BenchLM.ai