Best Long-Context AI Models

Ranked by context window size. Includes models with 32K+ token windows for document analysis, long-form reasoning, and large codebases.

Updated April 2026 · Source: AI Flash Report model database

#ModelContext windowInput/MInput modalities
#1 Gemini 3.1 Pro
Google
2M tokens $2.50 text, image, audio, video, PDF
#2 Kimi K2
Moonshot AI
2M tokens $0.15 text, image
#3 DeepSeek V3.2
DeepSeek
1M tokens $0.27 text
#4 Gemini 3 Pro
Google
1M tokens $2.50 text, image, audio, video, PDF
#5 Gemini 2.5 Flash
Google
1M tokens $0.30 text, image, audio, video
#6 Gemini 2.0 Flash
Google
1M tokens $0.10 text, image, audio, video
#7 Moonshot Kimi
Moonshot AI
1M tokens ¥12/M text, file
#8 Gemini 1.5 Pro
Google
1M tokens text, image, audio, video
#9 Claude Sonnet 4.6
Anthropic
500K tokens $3.00 text, image, PDF
#10 Claude Opus 4.5
Anthropic
500K tokens $15.00 text, image, PDF
#11 GPT-5.2
OpenAI
400K tokens $2.00 text, image, audio
#12 GPT-5.1
OpenAI
400K tokens $2.25 text, image, audio
#13 GPT-5
OpenAI
400K tokens $2.50 text, image, audio
#14 Mistral Large 3
Mistral
256K tokens $2.00 text, image
#15 GLM-5
Zhipu AI
200K tokens $0.11 text, image

More rankings