by Meta
Meta's open-weight frontier model — free to use, fine-tune, and deploy at any scale.
Llama 4 is Meta's open-source multimodal language model family, released in April 2025. The family includes Scout (10M token context), Maverick (400B parameters, 1M context), and Behemoth (288B active parameters) — all supporting text, image, and video understanding across 200+ languages. As open-weight models, they can be deployed and fine-tuned without API costs.
Llama 4 follows standard prompting conventions:
You are a helpful assistant specialized in code review.
Review this Python function for bugs, performance issues,
and style violations. Suggest fixes with code examples.
With Scout's 10M token window, you can process entire codebases or document libraries:
I'm providing our entire API documentation (500 pages).
Based on this, generate a migration guide from v2 to v3,
covering every breaking change.
Provide images alongside text for vision tasks:
Analyze this screenshot of our dashboard. Identify any UI/UX
issues and suggest improvements based on modern design principles.
| Parameter | Description |
|---|---|
| temperature | Randomness 0-2 |
| max_tokens | Maximum response length |
| top_p | Nucleus sampling threshold |
| system | System prompt for role/behavior |
Quick tips from the community about what works with Llama 4 right now.
Sign in to share a tip.
No tips yet. Add a tip for this model.