Skip to main content
DeepSeek·current

DeepSeek R1

DeepSeek's open reasoning model — o1-class math and code, fully open-weight

ProviderDeepSeek
Released2025-01-20
StatusCurrent
Context window66K tokens
PricingOpen
Input price$0.55/M
Output price$2.19/M
Capabilities
textcodereasoningmath
Hugging FaceView on HF ↗

Benchmarks

MMLU (knowledge)90.8
MATH (mathematics)97.3
GPQA (PhD science)71.5

As of 2025-02-01. Source

Loading…

Compare with similar models

DeepSeek R1Llama 3.1 405BDeepSeek V3Llama 3.3 70B
Context66K131K66K131K
MMLU90.888.688.586.0
HumanEval91.6
MATH97.373.590.277.0
GPQA71.5
PricingOpenOpenOpenOpen
Input $/M$0.55$0.27

Compare more models →