MODELS
Llama 4
Meta's open-weight multimodal model with a 1M-token context window.
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