Meta's 'Behemoth' Llama 4 model might still be months away

Last month, Meta hosted LlamaCon, its first ever generative AI conference. But while the event delivered some notable improvements for developers, it also felt a bit underwhelming considering how important AI is to the company. Now, we know a bit more about why, thanks to a new report in The Wall Street Journal.

According to the report, Meta had originally intended to release its "Behemoth" Llama 4 model at the April developer event, but later delayed its release to June. Now, it's apparently been pushed back again, potentially until "fall or later." Meta engineers are reportedly "struggling to significantly improve the capabilities" of the model that Mark Zuckerberg has called “the highest performing base model in the world.”

Meta has already released two smaller Llama 4 models, Scout and Maverick, and has also teased a fourth lightweight model that's apparently nicknamed "Little Llama." Meanwhile, the "Behemoth" model will have 288 billion active parameters and "outperforms GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro on several STEM benchmarks," the company said last month.

Meta has never given a firm timeline of when to expect the model. The company said last month that it was "still training." And while Behemoth got a few nods during the LlamaCon keynote, there were no updates on when it might actually be ready. That's probably because it could still be several months. Inside Meta there are apparently questions "about whether improvements over prior versions are significant enough to justify public release."

Meta didn't immediately respond to a request for comment. As the report notes, it wouldn't be the first company to run into snags as it races to release new models and outpace competitors. But the delay is still notable given the Meta's lofty ambitions when it comes to AI. Zuckerberg has made AI a top priority with Meta planning to spend as much as $72 billion on its AI infrastructure this year.

This article originally appeared on Engadget at https://www.engadget.com/ai/metas-behemoth-llama-4-model-might-still-be-months-away-221240585.html?src=rss https://www.engadget.com/ai/metas-behemoth-llama-4-model-might-still-be-months-away-221240585.html?src=rss
Created 2mo | May 15, 2025, 10:50:03 PM


Login to add comment

Other posts in this group

What the hell is going on with Subnautica 2?

If I had to describe the status of Subnautica 2 in just three words, it would be these: messy, messy, messy. That’s not to say the game itself is in terrible shape — this is actually a piv

Jul 18, 2025, 11:40:14 PM | Engadget
Netflix is already using generative AI in its original shows

Netflix admitted during its

Jul 18, 2025, 9:30:25 PM | Engadget
Microsoft unceremoniously kills off the Xbox Movies & TV store

Microsoft has rather abruptly closed down its Movies & TV app, which is accessible on Xbox and Windows PCs via the Microsoft Store. This allowed people to rent or buy movies or TV shows natively th

Jul 18, 2025, 7:10:23 PM | Engadget
Remedy lays out its plan to fix FBC: Firebreak, which includes improved onboarding

Remedy has announced plans to fix FBC: Firebreak and restore the good will of

Jul 18, 2025, 7:10:21 PM | Engadget
Meta says it won't sign the EU's AI code of practice

Meta said on Friday that it won't sign the European Union's new AI code of practice. The guidelines provide a framework for

Jul 18, 2025, 7:10:19 PM | Engadget
DuckDuckGo now allows you to filter out AI images in search results

DuckDuckGo is making it easier to wade through some of the AI slop that has taken over the internet in recent months. This week, the company

Jul 18, 2025, 4:50:07 PM | Engadget
Engadget Podcast: Ancestra director Eliza McNitt defends AI as a creative tool

Eliza McNitt is no stranger to new media. Her 2017 project, Fistful of Stars, was a fasc

Jul 18, 2025, 4:50:06 PM | Engadget