DeepSeek: DeepSeek V4 Pro preview

by DeepSeek intelligent open-weight coding reasoning agentic

DeepSeek V4 Pro is a Mixture-of-Experts model with 1.6 trillion total parameters and 1M-token context, scoring 9.0/10 and ranked #12 on LMRank. Launched as a preview in April 2026, it excels at coding, math, and agentic tasks, offering MIT-licensed weights and significantly lower cost than closed-source frontier models.

Input: $0.44/1M Output: $0.87/1M Context: 1,048,576 tokens Rank: #21

Compare this model Creative showcases