← Models / DeepSeek

DeepSeek R1 0528

671B MoE (37B active)

Upgraded R1 with 45-50% less hallucination, system prompt support. Available via third-party providers. Uses ~23K reasoning tokens per AIME question (doubled from R1's 12K).

reasoning OpenAI-compatible Function calling 131K context Released 2025-05

Pricing

Type USD / 1M tokens CNY / 1M tokens
Input $0.45 ¥4
Output $2.15 ¥16

vs Western Models

DeepSeek R1 0528 vs GPT-5.2 ($1.75/$14)
74% cheaper input 85% cheaper output
DeepSeek R1 0528 vs Claude Sonnet 4.6 ($3/$15)
85% cheaper input 86% cheaper output
DeepSeek R1 0528 vs Gemini 2.5 Pro ($1.25/$10)
64% cheaper input 79% cheaper output

Strengths

  • + reasoning
  • + math
  • + reduced hallucination vs R1
  • + supports system prompts

Watch Out For

  • ! content censorship

International Access

Direct access: Yes
Latency from US: 200-400ms
Registration: email
Payment: credit card

More from DeepSeek

Ready to try DeepSeek R1 0528?

Get your API key from the official platform.

Go to DeepSeek →