M
Llama 3.3 70B InstructOPEN
The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out).
Prices update automatically — checked daily against provider list prices.
- $0.71 per 1M tokens, assuming 3 input for 1 output the convention used by Artificial Analysis
- cheaper than 27% of the 191 open-weight models tracked here
- 10 of 12 Meta models tracked here, from cheapest
Benchmarks & performance
Independent benchmark scores for Llama 3.3 70B Instruct, measured by Artificial Analysis. Higher is better.
For the same budget, Llama 3.3 70B Instruct delivers 2.2 times more capability. Open weights also mean you can self-host it and pay nothing per token.
Models at this level cost $0.2 per 1M tokens (median of 26) — this one costs $0.71.
Cheaper and better on this index: GLM 5.3 Flash · Gemini 3.7 Flash (batch) · GPT-5.6 Luna and 81 more
About this model
Llama 3.3 70B Instruct is an open-weight AI model by Meta. You can download and self-host it for free; the prices below are hosted-API list prices, tracked daily, for when you prefer convenience over self-hosting.
Frequently asked questions
What is Llama 3.3 70B Instruct?
Llama 3.3 70B Instruct is an AI language model from Meta. It is open-weight: you can download it and run it on your own hardware, for free. It scores 9.3 on the Artificial Analysis intelligence index.
Is Llama 3.3 70B Instruct free?
The weights are free and open — you can self-host Llama 3.3 70B Instruct and pay nothing per token. If you prefer a hosted API, list prices are $0.71 per million input tokens and $0.71 per million output tokens.
What is Llama 3.3 70B Instruct good at?
Independent benchmarks from Artificial Analysis give it GPQA 49.8%, MMLU-Pro 71.3%, Humanity's Last Exam 3.6%, Long Context Reasoning 15.7%, LiveCodeBench 28.8%, SciCode 26%, MATH-500 77.3%, AIME 30%, AIME 2025 7.7%, IFBench 47.1%, τ²-Bench 26.6%, Terminal-Bench 4.9%, Terminal-Bench Hard 3%. It is particularly used for code generation.
How fast is Llama 3.3 70B Instruct?
It generates about 86.2 tokens per second, with a median 0.62s delay before the first token. Measured independently by Artificial Analysis.
Can I self-host Llama 3.3 70B Instruct?
Yes. Llama 3.3 70B Instruct has open weights, so you can download it and run it on your own GPU or server with tools like Ollama, vLLM or llama.cpp — with no per-token cost.