Model Directory› Meta› Llama 4 Maverick
Meta FlagshipOpen weights

Llama 4 Maverick

Meta's open-weight mixture-of-experts flagship, pairing native multimodality with a very low served price across third-party inference providers.

At a glance

Released
Feb 10, 2026
Provider
Meta
Context window
1,000,000 tokens
Max output
16,000 tokens
Modalities
Text + image → text
Knowledge cutoff
August 2025
API model ID
llama-4-maverick
License
Open weights (Llama 4 Community License)

Pricing

Standard
Input$0.20 / 1M
Output$0.80 / 1M

* Rates are from a third-party aggregator — verify before relying on them. A typical 1,500-token prompt with a 650-token answer costs ≈ $0.0008 on the standard tier. Run your own prompt through the calculator →

What's new vs the previous model

Compared with Llama 3.3 70B (December 2024)
  • Mixture-of-experts design lifts quality while keeping inference cheap
  • Native image understanding and much longer context
  • Larger total parameter footprint makes self-hosting heavier

Best use cases

Self-hosted assistants

Runs on your own infrastructure for data-sensitive chat and RAG.

Fine-tuning base

Serves as an open foundation for domain-specific adaptation.

Multimodal apps

Handles mixed text-and-image inputs without a proprietary API.

Strengths & limitations

Strengths

  • Open weights
  • Cheap to serve
  • Strong multimodal and long context

Limitations

  • Large footprint to host well
  • Community License is not fully unrestricted

How to test it

  1. No-code: open meta.ai and pick Llama 4 Maverick from the model switcher — no API key needed.
  2. Playground: sign in to llama.com (download the weights) and run prompts in the workbench, then copy the generated code.
  3. API: pick an inference host (Together, Fireworks, Groq, AWS Bedrock…), create a key and call its OpenAI-compatible endpoint, or read the docs.
# via an OpenAI-compatible inference host (Together, Fireworks, Groq, …)
curl https://<your-host>/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "llama-4-maverick", "messages": [{"role": "user", "content": "Explain tokenization in one paragraph."}]}'

Where to get access

Consumer app

Chat with Llama 4 Maverick in Meta AI (meta.ai, WhatsApp, Instagram, Messenger).

meta.ai ↗

Direct API

Pay-as-you-go access via llama.com (download the weights).

llama.com (download the weights) ↗

Also available

Hosted by Together, Fireworks, Groq, AWS Bedrock and others.

llama.meta.com ↗

Developer docs

API reference, SDKs and quickstarts for Meta.

llama.meta.com/docs ↗