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 · BenchGecko score 14.2. Top scores: MMLU-PRO (16.5%), BBH (HuggingFace) (14.2%), IFEval (13.4%).
HuggingFace MuSR (Multi-Step Reasoning). Tests multi-hop reasoning requiring chaining multiple facts together.
HuggingFace evaluation of MATH Level 5 problems. Competition math requiring advanced reasoning and proof construction.
HuggingFace MMLU-Pro. Harder version of MMLU with 10 answer choices instead of 4 and more challenging questions.
HuggingFace evaluation of GPQA (Graduate-Level Google-Proof Q&A). PhD-level science questions that cannot be easily searched.
- Typetext
- Context131K tokens (~66 books)
- ReleasedSep 2024
- LicenseOpen Source
- StatusActive
- Cost / Message~$0.000
Frequently Asked Questions
Key facts · as of 2026-07-11
- Llama 3.2 3B Instruct (free) by Meta. BenchGecko score 14.2, rank 293 of 312 scored models (normalized average of public benchmark scores).
- List price free input · free output per 1M tokens (as of 2026-07-11).
How to cite · data as of 2026-07-11
Llama 3.2 3B Instruct (free) · benchmarks, pricing and providers. BenchGecko, data as of 2026-07-11. https://benchgecko.ai/model/llama-3-2-3b-instruct-free
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