Meta Launches Llama 4 – New AI Model Supports 10 Million Tokens
Meta has unveiled its new Llama 4 model series, including the Scout and Maverick models. These new open AI models feature native multimodality and a context window of up to 10 million tokens.
What happened?
Meta has launched its new Llama 4 AI model family, which includes the Scout and Maverick versions. The models are based on a Mixture-of-Experts (MoE) architecture, with the Scout model featuring 17 billion active parameters distributed across 16 experts. One of the most significant specifications is the support for a context window of up to 10 million tokens, alongside native multimodal processing capabilities.
Key facts
| Modellserie | Llama 4 (Scout och Maverick) |
|---|---|
| Kontextfönster | Upp till 10 miljoner tokens |
| Arkitektur | Mixture-of-Experts (16 experter) |
| Aktiva parametrar (Scout) | 17 miljarder |
Why it matters
A context window of 10 million tokens is among the largest in the AI industry, enabling the processing of entire books, extensive codebases, or thousands of documents in a single request. By employing an MoE architecture, Meta is able to provide high performance and multimodality while maintaining computational efficiency.
Who is affected?
This launch is particularly relevant for AI developers, data analysts, and companies building large-scale applications for the analysis of long documents, source code, or complex log files. End users of personal AI assistants based on the Llama architecture will also benefit from the improved capacity to handle large volumes of data.
Impact on the EU
Llama 4 is being released with open weights, allowing European developers to download and run the models on their own infrastructure. This facilitates compliance with the EU's General Data Protection Regulation (GDPR) and the AI Act, as no sensitive data needs to be transmitted to servers outside the EU.
What else you should know
Meta has yet to publish full technical graphs for all benchmark tests; consequently, independent verification of the speed claims compared to GPT-4o and Mistral Large 4 is awaited from third-party evaluators.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
Påverkar det EU?
The link opens in a new window and leads to the publisher's own site.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.
AI-verktyg i artikeln
Topics
Get similar news straight to your inbox
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.
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 Launches Llama 4 – New AI Model Supports 10 Million Tok"