DeepSeek
Reasoning

DeepSeek-R1-Distill-70B

Distilled

R1 reasoning capability distilled into a Llama 3.1 70B base. The "open-weights o1-mini", strong STEM at single-machine deployable size.

DeepSeek-R1-Distill-70B is a reasoning AI model from DeepSeek. It costs $0.3500 per million input tokens and $0.8800 per million output tokens (blended $0.5090/M), with a 128,000-token context window.

RatesUSD per million tokens. Blend is 70% input, 30% output, our calculation.

INPUT
$0.3500/M
per million input tokens
OUTPUT
$0.8800/M
per million output tokens
BLENDED 70/30
$0.5090/M
unchanged since 3 May
CONTEXT
128,000
tokens

Hand-maintained from DeepSeek's published pricing; no check date on record.

What it is good at
  • Open-weights reasoning at 70B
  • Visible chain-of-thought
  • Strong math/code
Typical use cases
  • Self-hosted reasoning agents
  • STEM tutoring
  • Reasoning research

Benchmarks

vs. best public score
Graduate-level science questions, "Google-proof". Source: DeepSeek (self-reported), GPQA Diamond (pass@1), read 2026-11-10.
Hand-curated from each provider's published reports and public leaderboards. Methodology varies across sources, treat as directional rather than authoritative.

Also reported by the provider

Results on benchmarks outside the set we compare across models. They are listed for completeness and are deliberately excluded from the bars above and from every ranking on this site, because a score is only comparable against the same benchmark variant. Why benchmark scores do not compare →

  • MATH-500 (pass@1)94.5%self-reported by DeepSeek · no date on the source · R1 distilled-model table, Distill-Llama-70B, pass@1 from 64 responses per query, temperature 0.6, top-p 0.95, up to 32,768 tokens. Not comparable with the MATH column, so shown here and never ranked.

More from DeepSeek

See all 11 →

Frequently asked questions

How much does DeepSeek-R1-Distill-70B cost?

DeepSeek-R1-Distill-70B costs $0.3500 per million input tokens and $0.8800 per million output tokens, for a blended reference rate of $0.5090 per million tokens.

What is DeepSeek-R1-Distill-70B's context window?

DeepSeek-R1-Distill-70B supports up to 128,000 tokens of context in a single request.

What is DeepSeek-R1-Distill-70B best for?

DeepSeek-R1-Distill-70B is well suited to Open-weights reasoning at 70B, Visible chain-of-thought and Strong math/code.

Who makes DeepSeek-R1-Distill-70B?

DeepSeek-R1-Distill-70B is developed and served by DeepSeek. It was released in Jan 2025.

Terms used on this page