DeepSeek-V3.2
DeepSeek's latest flagship model matching GPT-5 performance with integrated tool-use thinking
DeepSeek-V3.2
DeepSeek • December 2025
Training Data
Up to December 2025
DeepSeek-V3.2
December 2025
Parameters
671 billion (37B active)
Training Method
Advanced MoE with Tool-use Training
Context Window
128,000 tokens
Knowledge Cutoff
December 2025
Key Features
Tool-use Integration • Thinking Modes • Agent Training • GPT-5 Competitive
Capabilities
Reasoning: Outstanding
Tool Use: Outstanding
Agent Tasks: Outstanding
What's New in This Version
First model to integrate thinking into tool-use, massive agent training data covering 1,800+ environments
DeepSeek's latest flagship model matching GPT-5 performance with integrated tool-use thinking
What's New in This Version
First model to integrate thinking into tool-use, massive agent training data covering 1,800+ environments
Technical Specifications
Key Features
Capabilities
Other DeepSeek Models
Explore more models from DeepSeek
DeepSeek-V3.2-Speciale
DeepSeek's competition-focused variant (EXPIRED Dec 15, 2025 - was temporary API-only release)
DeepSeek-V2
DeepSeek's flagship MoE model with exceptional efficiency
DeepSeek Coder V2
DeepSeek's specialized coding model with advanced programming capabilities
