This model is 19 months old. A newer version of it has since shipped: DeepSeek V4 Flash Vision Exp (2026-08-21). The details below are still accurate for this model — it just is not what we would recommend today.

DeepSeek R1 Distill Llama 70B

DeepSeek🇨🇳 China
active
Context window131K tokens
Input / 1M tokens$0.7
Output / 1M tokens$0.8

Version History

1.0major

DeepSeek released R1 Distill Llama 70B, a distilled 70B model combining Llama-3.3-70B-Instruct base architecture with fine-tuning from DeepSeek R1 outputs. Achieves 70% AIME 2024 pass@1 and 94.5% MATH-500 pass@1 at $0.70/$0.80 per 1M tokens.

Benchmark Scores

Full leaderboard →
81.5%
Hallucination Rate