DeepSeek V4.1 Flash
DeepSeek V4.1 Flash is served as DeepSeek-V4.1-Flash-1017 through the unchanged deepseek-flash API alias. It supports text and image input and a one-million-token context window; official weights are available under MIT.
Overview
DeepSeek released the 1017 revision on October 17, 2026, without changing the deepseek-flash API model name or token prices. The API documents up to 384K output tokens, thinking and non-thinking modes, tool calls, JSON output, Responses API and Anthropic-compatible access. Self-hosting requires substantial hardware; downloadable weights are not evidence of easy local deployment.
Model specifications
- Family
- DeepSeek V4
- Context
- 1,000,000 tokens
- Max output
- 384,000 tokens
- Input
- Text, Image
- Output
- Text
- Open weights
- Yes
- Capabilities
- Reasoning, Vision, Tool Use, Function Calling, Structured Output