DeepSeek-R1 is a LLM that excels in reasoning, code generation, and math tasks. It uses a chain-of-thought approach to break down complex problems into step-by-step solutions, making it particularly effective for logical reasoning and structured responses.

Metric	Value
Parameter Count	671 billion
Mixture of Experts	Yes
Active Parameter Count	37 billion
Context Length	128,000 tokens
Multilingual	No
Quantized*	Yes (FP8)

*Quantization is specific to the inference provider and the model may be offered with different quantization levels by other providers.