How much does Llama 3-70B cost?

Llama 3-70B is open source and can be self-hosted.

What benchmarks has Llama 3-70B been tested on?

Llama 3-70B has been evaluated on 18 benchmarks. Top scores: ARC AI2: 77.1, OpenBookQA: 76.8, MMLU: 72.4.

Is Llama 3-70B open source?

Yes, Llama 3-70B is open source.

How does Llama 3-70B compare to Gemma 2 2b It?

Llama 3-70B has an average score of 41.6 while Gemma 2 2b It scores 41.6. Gemma 2 2b It slightly outperforms Llama 3-70B overall. See full comparison →

Home/Models/Llama 3-70B

Llama 3-70B

Name: Llama 3-70B · benchmarks, pricing and providers
Creator: BenchGecko

by Meta · Released Jan 2024

Open Source

41.6

avg score

Rank #192

Compare

Better than 38% of all models

Context

N/A

Input $/1M

n/a

Output $/1M

n/a

Type

text

License

Open Source

Benchmarks

18 tested

Data as of 2026-06-22

About

Tested on 18 benchmarks · BenchGecko score 41.6. Top scores: ARC AI2 (77.1%), OpenBookQA (76.8%), MMLU (72.4%).

Capabilities

coding

5.0

#196 globally

reasoning

7.1

#194 globally

math

10.8

#252 globally

knowledge

45.0

#167 globally

general

25.2

#145 globally

language

33.2

#141 globally

Benchmark Scores

Compare All

Tested on 18 benchmarks · Ranked across 6 categories

Score Distribution (all 312 models)

0255075100

▲ You are here

codingCompare coding →

Cybench

Capture-the-flag cybersecurity challenges. Tests vulnerability analysis, reverse engineering, cryptography, and exploitation skills.

5.0·

reasoningCompare reasoning →

MUSR

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

7.1·

mathCompare math →

MATH level 5

Competition-level math from AMC, AIME, and olympiad problems. Level 5 is the hardest tier, requiring creative problem-solving.

22.6·

MATH Level 5

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

5.7·

OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

4.2·

Quick compare:

vs Gemma 2 2b It

vs Gemini 1.5 Flash (May 2024)

vs Llama 3 8B Instruct

Excellent (85+) Good (70-85) Average (50-70) Below (<50)

Model Family · Meta Llama 3

Llama 3 70B InstructApr 2024

30.6

$0.51/M in8Kctx9 benchmarks

Llama 3 8B InstructApr 2024

41.8+11.2

$0.14/M in(-0.37)8Kctx16 benchmarks

Llama 3-70BJan 2024

41.6-0.2

n/an/actx18 benchmarks

See the full Llama 3 family →

Similar Models

Gemma 2 2b It

Google DeepMind

41.6n/a

Gemini 1.5 Flash (May 2024)

Google DeepMind

41.7n/a

Llama 3 8B Instruct

Frequently Asked Questions

Llama 3-70B is an open-source text AI model by Meta, released in January 2024. It has an average benchmark score of 41.6.

Benchmarks

ARC AI2 OpenBookQA MMLU TriviaQA Winogrande

Meta · Provider Meta · Economy All Models Compare Models Pricing Developers · API

Key facts · as of 2026-06-22

Llama 3-70B by Meta. BenchGecko score 41.6, rank 192 of 312 scored models (normalized average of public benchmark scores).
List price n/a input · n/a output per 1M tokens (as of 2026-06-22).

How to cite · data as of 2026-06-22

Llama 3-70B · benchmarks, pricing and providers. BenchGecko, data as of 2026-06-22. https://benchgecko.ai/model/llama-3-70b

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

Llama 3-70B

Frequently Asked Questions

Related Models

Benchmarks

Related Pages

Key facts · as of 2026-06-22

How to cite · data as of 2026-06-22