Skip to content
Lanseringar· LaunchAvailable

Meta Releases Llama 3 – Open AI Model Challenges GPT-4

Meta is launching its highly anticipated open-source language model, Llama 3, along with its associated weights. Simultaneously, Meta AI is being made available directly in the browser without the need for a login.

By the Aheadline editorial team·5 aug. 2026·2 min read·Source: Entity-watch: Meta AIVerifierad signalAI-generated
Meta Releases Llama 3 – Open AI Model Challenges GPT-4
Meta Releases Llama 3 – Open AI Model Challenges GPT-4
Meta Releases Llama 3 – Open AI Model Challenges GPT-4
By · Verktygs- & infrastrukturreporter
Last updated

What happened?

Meta has officially released its new generation of open language models, Llama 3, in two sizes: 8 billion (8B) and 70 billion (70B) parameters. The models have been released in both pre-trained and instruction-tuned versions, with both the code and model weights made fully open. Additionally, Meta has made its AI assistant, Meta AI, available directly in the browser without requiring a login.

Key facts

Lanseringsdatum19 april 2024
Modellstorlekar8B och 70B parametrar
Träningsdata15 biljoner tokens
Hårdvara för träning24 000 GPU:er
Kontextfönster8k tokens

Why it matters

Llama 3 sets a new benchmark for open AI models, achieving State of the Art (SOTA) performance in their respective size classes. Performance in coding, reasoning, and mathematical problem-solving has been significantly improved, allowing even the smaller 8B model to outperform its predecessor, Llama 2 70B, in certain benchmarks.

Who is affected?

The launch primarily impacts AI developers, researchers, and companies looking to build their own applications on top of a powerful open language model without licensing costs. End-users are impacted by gaining free access to Meta AI on the web.

Impact on the EU

Llama 3 is released globally as open source and is available via Meta AI. Within the EU, however, access to the Meta AI interface may be affected by stricter privacy rules and regulations, such as the EU AI Act and GDPR.

What else you should know

Meta has trained Llama 3 on more than 15 trillion tokens in a custom cluster of 24,000 GPUs. The company is still developing an even larger model, Llama 3 400B, which is expected to be released later this year to challenge the most powerful closed-source models.

Frequently asked questions

Quick answers about this story

Vad har hänt?
Meta lanserade språkmodellen Llama 3 i versionerna 8B och 70B med öppen källkod samt gjorde Meta AI tillgängligt utan krav på inloggning.
När hände det?
Nyheten publicerades och modellerna släpptes den 19 april 2024.
Varför spelar det roll?
Det markerar ett stort framsteg för öppen källkod inom AI, då Llama 3 uppnår prestanda i klass med ledande stängda modeller som GPT-4.
Är Llama 3 tillgängligt i Sverige?
Ja, modellerna och koden finns tillgängliga globalt som öppen källkod för utvecklare och forskare.
Original source
Entity-watch: Meta AI·36kr.com

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

#Open Source#AI-benchmarking#Large Language Models (LLMs)#Models#Meta AI
[ 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 3 – Open AI Model Challenges GPT-4"