Models Model details

DeepSeek

DeepSeek: R1

deepseek/deepseek-r1

DeepSeek R1 is here: 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 & technical report. MIT licensed: Distill & commercialize freely!

Input
$0.7

/ 1M tokens

Output
$2.5

/ 1M tokens

Context window
64K

tokens

Limits

Context

64,000

tokens

Max output

16,000

25% of the context window

Parameters

684,489,845,504

Model facts

Author

DeepSeek

Tokenizer

DeepSeek

Modality

Text -> Text

Instruction format

Deepseek-r1

License

MIT

Knowledge cutoff

2024-07-31

Created

Jan 20, 2025

Hugging Face

deepseek-ai/DeepSeek-R1

Reasoning

Required

Publication status

Hugging Face: verifiedLicense: publishedParameter count: published

Modalities

Input · TextOutput · Text

Generation controls

Frequency penaltyInclude reasoningMax tokensPresence penaltyReasoningRepetition penaltyResponse formatSeedStopTemperatureTool choiceToolsTop kTop p

Capability scores

1.1
Agentic index

Better than 12% of models compared

24.6
Coding index

Better than 21% of models compared

11.4
Intelligence index

Better than 27% of models compared