Skip to content
Llama 4 logo

MODELS

Llama 4

Meta's open-weight multimodal model with a 1M-token context window.

metallamaopen source

Specs

Context window
1,000,000
Max output
8,192
Modalities
text, image
Tool use
Vision
Streaming
License
llama-4-community
Released
2025-04-01

Pricing

Llama 4 is Meta's latest open-weight family, handling text and images with native tool use and a 1M-token context window. It's aimed at teams who want to self-host a frontier-ish model without paying per-token API fees, and who can absorb the GPU cost. Output is capped at 8K tokens, which is tight for long-form generation. Released under the Llama 4 Community License — permissive for most uses, but not OSI-approved.

Editor's verdict

Pick Llama 4 when you need data sovereignty, customization, or predictable cost at scale — areas where GPT-5 or Claude can't follow you. On raw reasoning and coding it still trails the top closed models, and the 8K output cap will frustrate anyone generating long documents. The Community License is permissive enough for most products but read it before assuming it's "real" open source.

Reviews

No reviews yet. Be the first.

Last updated: 2026-04-29

We use cookies

Anonymous analytics help us improve the site. You can opt out anytime. Learn more