What Llama 4 Scout does
Llama 4 Scout is a 17B active parameter mixture-of-experts model with a 10 million token context window — the longest of any model in its class. Supports multimodal inputs and runs efficiently on a single H100 GPU. Meta open-source release enables full customization, fine-tuning, and on-premise deployment.
How Llama 4 Scout compares with its closest alternatives
4 of the 5 closest alternatives have a free plan, as Llama 4 Scout does. All 5 offer an API, as Llama 4 Scout does. Llama 4 Scout was founded in 2025; the alternatives were founded between 2023 and 2024, 5 of them earlier. Llama 4 Scout is based in Menlo Park, United States; all the 4 alternatives with a known headquarters are in the United States too.