← All modelsKey Specifications
Context Window164K
Max Output33K
TokenizerDeepSeek
Input Modalitiestext
Output Modalitiestext
Pricing
Input (per M tokens)$0.2100per million tokens
Output (per M tokens)$0.7900per million tokens
Capabilities
✓Function calling
✓Structured outputs
✓Unmoderated
About
DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context...
Supported parameters
frequency_penaltyinclude_reasoninglogit_biasmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_p