Board/Meta/Llama 4 Maverick
M

Llama 4 Maverick

Meta Superintelligence Labs·Apr 5, 2025N/A

Open weights

Legacy Llama-era MoE with a million-token window — ubiquitous fine-tuning base even after Meta's pivot to the Muse family.

Status
Generally available
License
Open weights · Custom
Context window
1M
Max output
33K
Input · $/M
$0.27
Output · $/M
$0.85
Cached input · $/M
—
Blended · $/M
$0.42
Parameters
400B MoE · 17B active
Inputs
Text + images

Benchmarks

This model has no published scores on the benchmarks we track yet. Specs and pricing below the fold are still fully sourced.

Reasoning & knowledge

  • GPQA Diamond

    no public score

  • MMLU-Pro

    no public score

  • Humanity's Last Exam

    no public score

  • ARC-AGI-2

    no public score

Coding

  • SWE-bench Verified

    no public score

  • SWE-bench Pro

    no public score

  • LiveCodeBench

    no public score

Agents & computer use

  • Terminal-Bench 2

    no public score

  • OSWorld 2.0

    no public score

  • BrowseComp

    no public score

Math

  • AIME 2026

    no public score

Known strengths

  • Huge fine-tune ecosystem
  • Million-token context

Primary sources

  • Llama models
Compare this model →

More from Meta

  • MMuse Glimmer 30BAug 10, 2026 · 256K ctx
  • MMuse Spark 1.1Jul 9, 2026 · 500K ctx