NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/R1 Distill Llama 70B
DeepSeek logo

R1 Distill Llama 70B

by DeepSeek · Released Jan 2025

Open Source
47.3
avg score
Rank #161
Compare
Better than 48% of all models
Context
8K tokens (~4 books)
Input $/1M
$0.80
Output $/1M
$0.80
Type
text
License
Open Source
Benchmarks
8 tested
Data as of
About

DeepSeek R1 Distill Llama 70B is a distilled large language model based on Llama-3.3-70B-Instruct, using outputs from DeepSeek R1. The model combines advanced distillation techniques to achieve high performance across...

Tested on 8 benchmarks · BenchGecko score 47.3. Top scores: IFEval (43.4%), MMLU-PRO (41.6%), BBH (HuggingFace) (35.8%).

Looking for similar performance at lower cost?
gpt-oss-120b scores 46.7 (99% as good) at $0.04/1M input · 95% cheaper
Capabilities
reasoning
13.3
#159 globally
math
30.7
#187 globally
knowledge
21.8
#264 globally
general
35.8
#93 globally
speed
13.7
#115 globally
language
43.4
#130 globally
Benchmark Scores
Compare All
Tested on 8 benchmarks · Ranked across 6 categories
Score Distribution (all 312 models)
0255075100
▲ You are here
reasoningCompare reasoning →
MUSR

HuggingFace MuSR (Multi-Step Reasoning). Tests multi-hop reasoning requiring chaining multiple facts together.

13.3·
mathCompare math →
MATH Level 5

HuggingFace evaluation of MATH Level 5 problems. Competition math requiring advanced reasoning and proof construction.

30.7·
knowledgeCompare knowledge →
MMLU-PRO

HuggingFace MMLU-Pro. Harder version of MMLU with 10 answer choices instead of 4 and more challenging questions.

41.6·
GPQA

HuggingFace evaluation of GPQA (Graduate-Level Google-Proof Q&A). PhD-level science questions that cannot be easily searched.

2.0·
Quick compare:Google DeepMind logovs Gemini 3.1 Flash LiteOpenAI logovs o1xAI logovs Grok-3 mini
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Similar Models
Google DeepMind logo
Gemini 3.1 Flash Lite
Google DeepMind
47.0$0.25/1M
OpenAI logo
o1
OpenAI
46.9$15.00/1M
xAI logo
Grok-3 mini
xAI
47.7n/a
Links
Info
DeepSeekPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@DeepSeek
Source Code
GitHubHugging Face
BenchGecko API
deepseek-r1-distill-llama-70b
Specifications
  • Typetext
  • Context8K tokens (~4 books)
  • ReleasedJan 2025
  • LicenseOpen Source
  • Statusexpired
  • Cost / Message~$0.002
Available On
DeepSeek logoDeepSeek$0.80
Categories
reasoningmathknowledgegeneralspeedlanguage
Learn More
transformeropen weightstokens
Share & Export
Tweet
Related Models
Google DeepMind logoGemini 3.1 Flash Lite
47.0
OpenAI logoo1
46.9
xAI logoGrok-3 mini
47.7
Google DeepMind logoGemma 4 26B A4B
47.8
R1 Distill Llama 70B is an open-source text AI model by DeepSeek, released in January 2025. It has an average benchmark score of 47.3. Context window: 8K tokens.

Related Models

Gemini 3.1 Flash Lite · Google DeepMindo1 · OpenAIGrok-3 mini · xAIGemma 4 26B A4B · Google DeepMindQwen3 32B · Alibaba Qwen

Benchmarks

IFEvalMMLU-PROBBH (HuggingFace)MATH Level 5Artificial Analysis — Quality Index

Related Pages

DeepSeek · ProviderDeepSeek · EconomyAll ModelsCompare ModelsPricingDevelopers · API

Key facts · as of 2026-09-28

  • R1 Distill Llama 70B by DeepSeek. BenchGecko score 47.3, rank 161 of 312 scored models (normalized average of public benchmark scores).
  • List price $0.80 input · $0.80 output per 1M tokens (as of 2026-09-28).

How to cite · data as of 2026-09-28

R1 Distill Llama 70B · benchmarks, pricing and providers. BenchGecko, data as of 2026-09-28. https://benchgecko.ai/model/deepseek-r1-distill-llama-70b

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP