DeepSeek V4-Pro: Open Models, OpenAI API, Lower Off-Peak Pricing
DeepSeek V4-Pro matches top open models, nears GPT-5.6. New features: OpenAI Responses API, three reasoning levels. Pricing: peak/off-peak with 50% off-peak discount. Available Aug 16, 2026.
DeepSeek has released DeepSeek-V4-Pro.
It is approximately at the level of the best open models — Kimi k3 and GLM-5.2, and is already approaching GPT-5.6 Sol and Fable 5.
Among the technical updates — native support for OpenAI Responses API and one-click Codex configuration. The model names in the API remain unchanged. For V4 Pro and Flash, three reasoning levels have also appeared: low, high, and maximum.
At the same time, pricing is also changing: peak and off-peak rates are being introduced for the V4 line, with prices being 50% lower during off-peak hours.
The new rates will take effect on August 16, 2026 at 16:00 UTC.
V4-Pro is already available via API, as well as in "Expert Mode" in the app and web version.
Comments
Weekly digest
The best of vibe coding, AI agents and open source — once a week, no spam.
Related articles
All articlesGrok 4.6 Released: Matches GPT-5.6, Costs Half as Much
xAI's Grok 4.6 matches GPT-5.6 Sol in AA Intelligence Index, surpasses in CursorBench and FrontierCode. Priced at $2/M input, $6/M output. Available in Grok Build, Cursor, API.
Claude Sonnet 5 Released: New Default Model, Lower Prices and Limits
Claude Sonnet 5 is now the default for Free and Pro. Quality near Opus 4.8, cheaper pricing through summer, but usage limits were lowered.
Grok 4.5 Release: Cheaper Than GPT and Claude With Top Quality
Grok 4.5 matches or beats GPT and Claude while costing less. See token pricing vs GPT-5.6 and Opus 4.8 and why it pressures OpenAI and Anthropic.