deepseek/deepseek-r1-0528
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...
| Provider | In $/M | Out $/M | Context | Uptime |
|---|---|---|---|---|
| DeepInfrafp4 | $0.500 | $2.15 | 164K | 100% |
| SiliconFlowfp8 | $0.500 | $2.18 | 164K | 99.7% |
| AtlasCloudfp8 | $0.550 | $2.15 | 131K | 99.8% |
| StreamLake | $0.571 | $2.29 | 128K | 100% |
| Novitafp8 | $0.700 | $2.50 | 164K | — |
R1 0528 costs $0.500 per million input tokens and $2.15 per million output tokens via OpenRouter, making it 164th cheapest of 298 paid models.
R1 0528 scores 27.1 on the Artificial Analysis Intelligence Index, ranking 86th of 180 benchmarked models, with a GPQA Diamond score of 81%.
R1 0528 generates around 30 tokens per second with 627ms time-to-first-token (p50), the 242nd fastest tracked model.
R1 0528 supports a 164K-token context window and can output up to 33K tokens. It accepts text input.