gpt.buzz
Sign in

Compare models

Pick up to 4models. Specs render side-by-side. Share the URL — it's stateless.

SelectedDeepSeek logoDeepSeek-V4-Pro×Google logoGemini 2.5 Pro×Anthropic logoClaude 4.6 Sonnet×Clear all
 
DeepSeek logoDeepSeek-V4-Pro

DeepSeek

Google logoGemini 2.5 Pro

Google

Anthropic logoClaude 4.6 Sonnet

Anthropic

VendorDeepSeekGoogleAnthropic
FamilyDeepSeekGeminiClaude
Release date2026-04-222025-03-25
Context window1,000,000 tokens2,000,000 tokens200,000 tokens
Parameters1.6T (49B active)
Modalitytexttext, vision, audio, videotext, vision
LicenseMITproprietaryproprietary
Sourceopen weightsproprietaryproprietary
DescriptionDeepSeek's flagship open-weight MoE. 1.6T parameters with 49B activated, 1M-token context, and a hybrid attention scheme (CSA + HCA) that delivers long-context inference at ~27% of V3.2's FLOPs.Predecessor to Gemini 3 Pro.
Links

Add a model