DeepSeek has moved DeepSeek-V4-Pro to general availability, presenting it as a major upgrade for agent-based workloads with notable gains in production scenarios.

The V4 lineup introduces flexible reasoning effort: DeepSeek-V4-Pro and V4-Flash can run at a low setting for simple tasks, high for everyday agent workflows, and max for complex problems. The platform also adds native support for the OpenAI Responses API, with an optimized one-click setup for Codex.

V4-Pro is available in the app and on the web via an 'Expert Mode' option. For API users, the model names remain unchanged; configuration details are documented in the API docs.

Alongside the release, DeepSeek is restructuring its API pricing with peak and off-peak rates. Off-peak usage is priced 50% lower than peak, aiming to encourage more flexible workload scheduling. The new pricing takes effect at 16:00 UTC on August 16, 2026.