Open-source MoE language model
Official DeepSeek release of a Mixture-of-Experts language model with 671B total parameters, 37B activated parameters, and 128K context length.
Prices are quoted per 1M tokens in the DeepSeek-V3 launch announcement.
Official DeepSeek API Docs launch post for DeepSeek-V3 lists token pricing at $0.27 per 1M input tokens (cache miss), $0.07 per 1M input tokens (cache hit), and $1.10 per 1M output tokens, effective from Feb 8 onward.
