Superseded
This post names retired API model IDs (deepseek-v4-flash and related strings). Use deepseek-flash for V4.1 Flash today. See the model name migration guide and the V4.1 Flash release note.
DeepSeek-V4
The DeepSeek API now supports V4-Pro and V4-Flash, available via both the OpenAI ChatCompletions interface and the Anthropic interface.
DeepSeek's change log says: The DeepSeek API now supports V4-Pro and V4-Flash, available via both the OpenAI ChatCompletions interface and the Anthropic interface. To access the new models, the base_url remains unchanged, and the model parameter should be set to deepseek-v4-pro or deepseek-v4-flash.
Who it affects: API users selecting deepseek-v4-pro or deepseek-v4-flash.
What it means: Check the official change log entry and update model names or integration tests if your app pins a specific model version.
Official sources
Related pages: /models · /api-pricing · /guides/model-name-migration · /guides/off-peak-pricing · /news
Context for developers
DeepSeek-V4 introduced deepseek-v4-pro and deepseek-v4-flash on the existing https://api.deepseek.com host with both OpenAI Chat Completions and Anthropic-compatible clients. The same announcement scheduled legacy deepseek-chat / deepseek-reasoner discontinuation for July 2026.
Later 2026 entries supersede the Flash string with deepseek-flash for V4.1 — read the V4.1 release note before pinning model names in new code. Peak/off-peak billing arrived in the same era; see /api-pricing.
Independent resource, not affiliated with DeepSeek.