Curated model list
Models with the longest context windows in 2026
Long context has gone from research curiosity to production necessity. Modern Agent stacks chew through 100K+ tokens routinely. These are the models that handle it without losing the plot in the middle.
Llama 4 Scout
Meta10,000,000-token context. Meta's long-context offering.
open sourceGemini 3 Pro
Google2,000,000-token context. Google DeepMind's flagship multimodal model
Gemini 2.5 Pro
Google2,000,000-token context. Predecessor to Gemini 3 Pro.
DeepSeek-V4-Pro
DeepSeek1,000,000-token context. DeepSeek's flagship open-weight MoE
open sourceDeepSeek-V4-Flash
DeepSeek1,000,000-token context. Smaller, faster sibling to DeepSeek-V4-Pro
open sourceGPT-4.1
OpenAI1,000,000-token context. OpenAI's long-context offering.
Claude 4.7 Opus
Anthropic1,000,000-token context. Anthropic's most capable model, optimized for complex reasoning and coding
Gemini 2.5 Flash
Google1,000,000-token context. Google's long-context offering.
Want the rest? Browse the full model catalog, or build a side-by-side comparison.