DeepSeek R1 0528
DeepSeekFlagship
ThinkingTool UseStructured Output
About this model
May 28th update to the [original DeepSeek R1](/deepseek/deepseek-r1) Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens. It's 671B parameters in size, with 37B active in an inference pass. Fully open-source model.
Performance Tier
Flagship
DeepSeek R1 0528 is a flagship model from DeepSeek : the most capable in their lineup.
Best-in-class model from this provider. Highest performance across benchmarks, ideal for demanding tasks.
Pricing
This model is included in Elosia plans| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.450 |
| Output (completion) | $2.15 |
| Cache read | $0.225 |
Capabilities
Context Length164K
Max Output Tokens66K
TokenizerDeepSeek
Inputtext
Outputtext
Release DateMay 28, 2025
Benchmarks
General Intelligence
MMLU
90.5%
MMLU-Pro
85%
GPQA Diamond
87.5%
Mathematics
MATH-500
97.3%
AIME 2025
87.5%
Programming
HumanEval
88.5%
SWE-bench Verified
57.6%
LiveCodeBench
73.3%
Reasoning
IFEval
84%
Humanity's Last Exam
17.7%
Recommended Use Cases
MathematicsResearchAnalysisCoding
Strengths
- Exceptional mathematical reasoning (MATH-500 97.3%, AIME 96.3%)
- Strong science reasoning rivaling top proprietary models (GPQA 87.5%)
- Open-weight reasoning model with transparent chain-of-thought
- Excellent value — top-tier reasoning at open model pricing
Limitations
- Slower responses due to extended thinking process
- Less capable on software engineering tasks (SWE-bench 57.6%)
- Creative writing and conversation less natural than chat-optimized models
Resources
This model may use your data for training
