DeepSeek retires legacy API aliases as V4 goes GA with new pricing
DeepSeek-V4 Pro and Flash replace old names; aliases deprecated in 3 months.
Deep Dive
DeepSeek API now supports V4-Pro and V4-Flash via both OpenAI ChatCompletions and Anthropic interfaces. The legacy model aliases deepseek-chat and deepseek-reasoner will be discontinued on July 24, 2026. During the transition period, these aliases point to V4-Flash's non-thinking and thinking modes respectively.
Key Points
- DeepSeek-V4 models (deepseek-v4-pro, deepseek-v4-flash) now generally available via OpenAI and Anthropic API interfaces.
- Legacy aliases deepseek-chat and deepseek-reasoner to be discontinued on July 24, 2026; currently point to V4-Flash modes.
- Developers must update model parameters to new names to avoid service disruption; pricing applies to V4 models.
Why It Matters
Forces API migration by July 2026; V4 brings native thinking/non-thinking modes with new pricing structure.