Models

OpenAI

GPT-5.2

256K context Weekly +8.1%
Use model
OpenAI

GPT-5 mini

256K context Weekly +14.2%
Use model
Anthropic

Claude 4 Sonnet

200K context Weekly +9.5%
Use model
Anthropic

Claude 4 Opus

200K context Weekly −1.2%
Use model
Google

Gemini 3.1 Pro

2M context Weekly +6.8%
Use model
Google

Gemini 3.1 Flash

1M context Weekly +19.3%
Use model
Meta

Llama 4 405B

128K context Weekly +7.2%
Use model
Mistral AI

Mistral Large 2

128K context Weekly +4.1%
Use model
DeepSeek

DeepSeek V4

128K context Weekly +24.1%
Use model