M
ModelMeta
</> API
499 Models·19 Providers·Updated 2h ago
All systems operational
Back to models

DeepSeek-Reasoner

Flagship modelRecommended

DeepSeek-V3.2 in thinking mode. 128K context with extended reasoning capabilities, max 64K output tokens.

Context window64,000 tokens
Max output tokens64,000 tokens

Pricing

TypeInputOutput
Standard$0.56 / 1M tokens$2.22 / 1M tokens

Parameters

ParameterDefaultRange
Temperature10 to 2
Top P10 to 1
Frequency penalty0-2 to 2
Presence penalty0-2 to 2
Max stop sequences4

Supported features

Capabilities

  • Streaming
  • Function Calling
  • Structured Output
  • Json Mode
  • Reasoning
  • System Prompt

Endpoints

  • Chat Completions

Use cases

  • text-chat
  • function-calling
  • reasoning

Model specifications

Model IDdeepseek-reasoner
Familydeepseek-v3
Access typeClosed
Input modalitiestext
Output modalitiestext