How much does Llama 3.2 3B Instruct (free) cost?

Llama 3.2 3B Instruct (free) costs $0.00 per million input tokens and $0.00 per million output tokens. For a typical conversation (~2,000 tokens), that's approximately $0.000 per message.

What benchmarks has Llama 3.2 3B Instruct (free) been tested on?

Llama 3.2 3B Instruct (free) has been evaluated on 6 benchmarks. Top scores: MMLU-PRO: 16.5, BBH (HuggingFace): 14.2, IFEval: 13.4.

Is Llama 3.2 3B Instruct (free) open source?

Yes, Llama 3.2 3B Instruct (free) is open source.

How does Llama 3.2 3B Instruct (free) compare to Qwen2.5 Coder 0.5B Instruct?

Llama 3.2 3B Instruct (free) has an average score of 14.7 while Qwen2.5 Coder 0.5B Instruct scores 14.3. Llama 3.2 3B Instruct (free) outperforms Qwen2.5 Coder 0.5B Instruct overall. See full comparison →

Home/Models/Llama 3.2 3B Instruct (free)

Llama 3.2 3B Instruct (free)

Name: Llama 3.2 3B Instruct (free)
Author: Meta

by Meta · Released Sep 2024

Open Source

14.7

avg score

Rank #216

Compare

Better than 7% of all models

Context

131K tokens (~66 books)

Input $/1M

Free

Output $/1M

Free

Type

text

License

Open Source

Benchmarks

6 tested

Data updated today

About

Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization. Designed with the latest transformer architecture, it...

Tested on 6 benchmarks with 8.7% average. Top scores: MMLU-PRO (16.5%), BBH (HuggingFace) (14.2%), IFEval (13.4%).

Capabilities

reasoning

3.8

#166 globally

math

1.9

#205 globally

knowledge

9.4

#211 globally

language

13.4

#150 globally

general

14.2

#47 globally

Benchmark Scores

Compare All

Tested on 6 benchmarks · Ranked across 5 categories

Score Distribution (all 233 models)

0255075100

▲ You are here

reasoningCompare reasoning →

MUSR

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

3.8—

mathCompare math →

MATH Level 5

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

1.9—

knowledgeCompare knowledge →

MMLU-PRO

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

16.5—

GPQA

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

2.4—

Quick compare:

vs Qwen2.5 Coder 0.5B Instruct

vs Llama 4 Scout

vs DeepSeek R1 Distill Qwen 1.5B

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

Model Family · Meta Llama 3.2

Llama 3.2 11B Vision InstructSep 2024

$0.24/M in131Kctx

Llama 3.2 1B InstructSep 2024

14.5+14.5

$0.03/M in(-0.22)60Kctx(-71K)7 benchmarks

Llama 3.2 3B InstructSep 2024

24.2+9.7

$0.05/M in(+0.02)80Kctx(+20K)7 benchmarks

Llama 3.2 3B Instruct (free)Sep 2024

8.7-15.5

$0.00/M in(-0.05)131Kctx(+51K)6 benchmarks

Llama 3.2 90BJan 2024

36.1+27.4

N/AN/Actx6 benchmarks

See the full Llama 3.2 family →

Similar Models

Qwen2.5 Coder 0.5B Instruct

Alibaba

14.3TBD

Llama 4 Scout

Frequently Asked Questions

Llama 3.2 3B Instruct (free) is an open-source text AI model by Meta, released in September 2024. It has an average benchmark score of 14.7. Context window: 131K tokens.

Benchmarks

MMLU-PRO BBH (HuggingFace)IFEval MUSR GPQA

Meta · Provider Meta · Economy All Models Compare Models Pricing Developers · API Context Window · Glossary

Llama 3.2 3B Instruct (free)

Frequently Asked Questions

Related Models

Benchmarks

Related Pages