Skip to content
Meta· LaunchAvailable

Meta releases Llama 4 Scout with Mixture-of-Depths network architecture

Meta has launched Llama 4 Scout, an open foundation model with 105 billion parameters based on a Mixture-of-Depths architecture. The model supports a 1 million token context window.

By the Aheadline editorial team·13 okt. 2026·2 min read·Source: Entity-watch: Meta AIVerifierad signalAI-generated
Meta releases Llama 4 Scout with Mixture-of-Depths network architecture
Meta releases Llama 4 Scout with Mixture-of-Depths network architecture
Meta releases Llama 4 Scout with Mixture-of-Depths network architecture
By · Verktygs- & infrastrukturreporter
Last updated
Vad betyder det för mig?

What happened?

Meta's research team, FAIR, has announced Llama 4 Scout, an open AI model with a total of 105 billion parameters. The model uses a Mixture-of-Depths (MoD) architecture, where only 24 billion parameters are activated per token, and features a built-in context window of 1 million tokens.

Key facts

ModellnamnLlama 4 Scout
Totala parametrar105 miljarder
Aktiva parametrar per token24 miljarder
Kontextfönster1 000 000 tokens
Lanseringsdatum11 oktober 2026

Why it matters

By using Mixture-of-Depths, Meta dramatically streamlines computational capacity (FLOPs). This allows Llama 4 Scout to achieve performance parity with leading closed models in coding, mathematics, and long-context analysis, while running efficiently on workstations equipped with two graphics cards.

Who is affected?

The model is primarily aimed at AI developers, researchers, and companies looking to run advanced reasoning and coding models locally on their own hardware or in private clouds, without relying on closed API services.

Impact on the EU

The model has been released under an open licence and is available globally for download and local operation, making it useful for European developers subject to stringent data sovereignty requirements and GDPR.

What else you should know

Llama 4 Scout is the first model in Meta's Llama 4 series to introduce dynamic depth routing at scale. By reducing the number of active parameters per token, the model enables more cost-effective local AI agent applications.

Frequently asked questions

Quick answers about this story

Vad har hänt?
Meta har släppt Llama 4 Scout, en öppen AI-modell baserad på Mixture-of-Depths-arkitektur med 1 miljon tokens i kontextfönster.
När hände det?
Modellen offentliggjordes den 11 oktober 2026 av Metas AI-forskningsgrupp FAIR.
Varför spelar det roll?
Den gör det möjligt att köra avancerade resonemangs- och kodningsmodeller på vanliga dubbla grafikkort istället för stora serverkluster.
Vilka berörs av släppet?
AI-utvecklare och företag som behöver köra storskalig språkmodellering lokalt med hög beräkningseffektivitet.
Original source
Entity-watch: Meta AI·worldprogramming.org

The link opens in a new window and leads to the publisher's own site.

Verifierad signal

Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.

AI-verktyg i artikeln

Topics

#Meta
[ STAY UP TO DATE ]

Get similar news straight to your inbox

No affiliate linksCancel anytimeGDPR-friendly
[ Frequency ]
[ What do you want to read about? ]

You'll receive updates on 2 topics.

The reader's room

Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.

Sign in to submit a comment or question.

Loading comments…
How this affects you

Read the article through your role

  • Decide whether this affects strategy over 6–12 months or is just noise.
  • Discuss with leadership: do we own the right question or does ownership need to move?
  • Ask: what risk are we taking by NOT acting on this this quarter?

Generated angle — not editorial analysis of "Meta releases Llama 4 Scout with Mixture-of-Depths network a"