◑

Llama 4

by Meta

4.5
Benchmark RatingLaunched 2025Free / open weights

Llama 4 is Meta's open model family spanning Scout (efficient), Maverick (balanced) and Behemoth (frontier). Run them locally, fine-tune on your data, or use via partner clouds — no vendor lock-in.

Open weightsFine-tunableMulti-modal
Visit Site

Pricing Tiers

Click a tier to see full details

Open weights — $0

Subscribe

Self-host, free

Plan: Open weightsPrice: $0Tool: Llama 4

Capabilities & Specs

Best for
Open & self-hosted LLM
Context window
131K
Image generation
Voice mode
Code execution
Web browsing
Offline / open weights
Public API
Starting price
Free / open

Pros

  • Fully open weights
  • Runs on consumer GPUs (Scout)
  • Huge community

Cons

  • Self-hosting needs infra
  • No first-party chat app

Available Models

ModelContextPriceNote
Llama 4 Maverick1MFreeBalanced open model
Llama 4 Scout10MFreeEfficient, runs locally
Llama 3.3 70B128KFreePrevious flagship
Llama 3.1 8B128KFreeSmall + fast

Tasks Llama 4 Covers

WritingCodingAgents