DeepSeek R1 (0528) is an updated checkpoint of DeepSeek's R1 reasoning model with improved chain-of-thought quality and benchmark performance. It targets the same complex math, coding, and logic reasoning use cases as the original R1.
Cost rate
0.5x
Context
164K
Released
May 28, 2025
Input
Text
Output
Text
Support
ReasoningTool callingStructured outputs
Performance
Median latency and throughput measured across recent requests.
Throughput31 tok/s
Latency3.56s
Best at
The categories where this model ranks highest.
Medicine & Healthcare
#129 · top 33%
Entertainment, Sports, & Media
#134 · top 32%
Software & IT Services
#136 · top 32%
Mathematical
#150 · top 37%
| Domain | Score | Rank |
|---|---|---|
| Software & IT Services | 1456 | #136 |
| Writing, Literature, & Language | 1388 | #151 |
| Life, Physical, & Social Science | 1426 | #157 |
| Entertainment, Sports, & Media | 1388 | #134 |
| Business, Management, & Finance | 1407 | #158 |
| Mathematical | 1415 | #150 |
| Legal & Government | 1415 | #155 |
| Medicine & Healthcare | 1443 | #129 |
| Category | Win rate | Rank |
|---|---|---|
| Data Visualization | 60.9% | #68 |
| SVG | 48.7% | #87 |
| 3D | 53.3% | #98 |
| Website | 52.7% | #101 |
| Code Categories | 52.6% | #102 |
| UI Component | 54.9% | #105 |
| Game Development | 49.3% | #106 |
Latest news about R1 0528
Recent articles and community posts, linked to their original sources.
You Need to Learn How to Use Deepseek AI Right NowTech Meets East6,830
I Tried to Clone Claude Code for FREE with REA + DeepSeekWeb Tech Knowledge1,836
DeepSeek + REA (Reverse Engineer Anything) is INSANE!Julian Goldie SEO2,431
Deepseek’s MASSIVE AI Breakthrough!Dr. Reed Household Secrets7,935- Show HN: DeepSeek V3 in ~200 lines of pure JaxHacker News27630
DeepSeek V4 Pro 0813Hacker News251