M
Llama 4 ScoutOPEN
Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B.
Prices update automatically — checked daily against provider list prices.
- $0.17 per 1M tokens, assuming 3 input for 1 output the convention used by Artificial Analysis
- cheaper than 65% of the 191 open-weight models tracked here
- 5 of 12 Meta models tracked here, from cheapest
Benchmarks & performance
Independent benchmark scores for Llama 4 Scout, measured by Artificial Analysis. Higher is better.
For the same budget, Llama 4 Scout delivers 5.4 times more capability. Open weights also mean you can self-host it and pay nothing per token.
Models at this level cost $0.23 per 1M tokens (median of 24) — this one costs $0.17.
Cheaper and better on this index: GLM 5.3 Flash · DeepSeek V4 Flash 0731 · DeepSeek V4 Flash 0423 and 20 more
About this model
Llama 4 Scout 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 4 Scout?
Llama 4 Scout 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 10.3 on the Artificial Analysis intelligence index.
Is Llama 4 Scout free?
The weights are free and open — you can self-host Llama 4 Scout and pay nothing per token. If you prefer a hosted API, list prices are $0.11 per million input tokens and $0.34 per million output tokens.
What is Llama 4 Scout good at?
Independent benchmarks from Artificial Analysis give it GPQA 58.7%, MMLU-Pro 75.2%, Humanity's Last Exam 3.8%, Long Context Reasoning 30.3%, LiveCodeBench 29.9%, SciCode 17%, MATH-500 84.4%, AIME 28.3%, AIME 2025 14%, IFBench 39.5%, τ²-Bench 15.5%, τ-Bench Banking 3.3%, Terminal-Bench 3.7%, Terminal-Bench Hard 1.5%. It is particularly used for code generation.
How fast is Llama 4 Scout?
It generates about 137.4 tokens per second, with a median 0.56s delay before the first token. Measured independently by Artificial Analysis.
Can I self-host Llama 4 Scout?
Yes. Llama 4 Scout 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.