AI Launch Profile

The Llama 4 herd

Meta introduced the Llama 4 herd with Scout and Maverick as open-weight natively multimodal mixture-of-experts models, plus a Behemoth preview.

Editorial image for The Llama 4 herd AI launch

At a glance

Launch Snapshot

Company
Meta
Launch date
April 5, 2025
Launch type
Open-Source Release
Category
AI Open-Weight Models
Audience
Developers, Enterprises, Researchers
Pricing
Weights and partner platform access are available; actual cost depends on hosting, inference provider, and hardware requirements.
Free plan
Yes
API
Yes
Open weights/source
Yes

Verification & Sources

Status
Verified
Source links
2
Freshness
Needs recheck: verified June 8, 2026
Last verified
June 8, 2026
Last updated
June 11, 2026

Key source checks

Suggest a correction

Form submissions, correction notes, score details, URLs, and analytics events may be stored for editorial review, spam prevention, product improvement, and follow-up. Do not submit secrets, unreleased financials, private customer data, or regulated personal data through these forms.

Kingy Launch Score

7.2 / 10 · Solid

One earned credibility score, computed from cited evidence — not a placeholder. How the Kingy Launch Score works

Why this score

  • Source & verification 7.0/10 — The Meta announcement plus the Hugging Face model org — two first-party sources. ai.meta.com
  • Product evidence 8.0/10 — Open-weight natively multimodal mixture-of-experts models (Scout and Maverick) plus a Behemoth preview, with partner-platform access. huggingface.co
  • Significance & novelty 7.0/10 — The record frames it as a major open-weight release that reset Meta's direction toward native multimodality and mixture-of-experts architecture. ai.meta.com
  • Traction signals not scored — insufficient sourced evidence
  • Offer clarity 6.5/10 — Weights and partner-platform access are available; cost depends on hosting, inference provider, and hardware. ai.meta.com

Evidence checked 2026-07-11

Kingy AI Take

Meta introduced the Llama 4 herd with Scout and Maverick as open-weight natively multimodal mixture-of-experts models, plus a Behemoth preview (ai.meta.com). The release shifts Meta's open-model direction toward native multimodality and MoE for developers and researchers on open weights. Benchmark claims, real-world quality, and license constraints need ongoing independent validation, so test Scout and Maverick on your own workloads first.

Who it is for

Developers, researchers, and enterprises building multimodal applications, long-context systems, and open-weight model products.

Traction notes

Major open-weight release because Llama moved to native multimodality and mixture-of-experts architecture.