Compare models

Pick up to four models to compare specs and prices side by side. The best context window and lowest price in each row are marked.

Attribute DeepSeekMath-V2DeepSeek
Overview
Developer DeepSeek
Released Nov 27, 2025
Type Text LLM
Weights Open weights
Status Active
License Apache License 2.0
Size and context
Architecture MoE
Parameters 671B
Active parameters 37B
Context window -
Max output -
Knowledge cutoff -
Price per 1M tokens (USD)
Input, current -
Output, current -
Input, at launch -
Output, at launch -

Add another model above to compare.