Home Models Compare Local Models Pricing Scorecards Evals OpenClaw Methodology

Long-context guide

Best Long-Context Models

If you process giant documents, repos, transcripts, or agent memory buffers, context window stops being a vanity metric and becomes the whole damn game.

ModelContextBest forInput $/M
Qwen 3 Max
Alibaba
262KMultilingual$1.2
Kimi K2.5
Moonshot AI
256KVisual coding$0.6
GLM-5
Zhipu AI
205KBilingual (CN/EN)$0.5
Claude Opus 4.6
Anthropic
200KComplex reasoning$5
Claude Sonnet 4.6
Anthropic
200KBalanced performance$3
GPT-5.3 Codex
OpenAI
200KCoding-focused tasks$3
MiniMax M2.5
MiniMax
196KReal-world productivity$0.3
Llama 4 (405B)
Meta
128KSelf-hosted$2
DeepSeek V3
DeepSeek
128KBudget coding$0.27
GPT-5.2
OpenAI
128KGeneral-purpose$1.75
Mistral Large 3
Mistral
128KEuropean compliance$2
GPT-OSS-120B
OpenAI
128KSelf-hosted$0
Grok 4.1 Fast
xAI
2MLong context$0.2
GPT-5.4
OpenAI
1.05MCoding$2.5
Gemini 3.1 Pro
Google
1MMultimodal tasks$1.25

Fast picks