DeepSeek-R1-0528
DeepSeek's upgraded reasoning model with 87.5% AIME accuracy and significantly reduced hallucinations
DeepSeek-R1-0528
DeepSeek • May 2025
Training Data
Up to May 2025
DeepSeek-R1-0528
May 2025
Parameters
671 billion (37B active)
Training Method
Enhanced Reinforcement Learning
Context Window
128,000 tokens
Knowledge Cutoff
May 2025
Key Features
Enhanced Reasoning • System Prompts • JSON Output • Function Calling • Reduced Hallucinations
Capabilities
Reasoning: Outstanding (87.5% AIME)
Math: Outstanding
Function Calling: Outstanding
What's New in This Version
Major upgrade: 87.5% AIME accuracy (vs 70%), deeper thinking (23K vs 12K tokens), 45-50% fewer hallucinations
DeepSeek's upgraded reasoning model with 87.5% AIME accuracy and significantly reduced hallucinations
What's New in This Version
Major upgrade: 87.5% AIME accuracy (vs 70%), deeper thinking (23K vs 12K tokens), 45-50% fewer hallucinations
Technical Specifications
Key Features
Capabilities
Other DeepSeek Models
Explore more models from DeepSeek
DeepSeek-V4-Pro
DeepSeek's frontier MoE flagship closing the gap with leading proprietary models on reasoning and agentic coding
DeepSeek-V4-Flash
DeepSeek's smaller, fast variant of V4 — same architecture at a fraction of the cost and latency
DeepSeek-V3.2
DeepSeek's latest flagship model matching GPT-5 performance with integrated tool-use thinking
