Meta confirms plans to start rolling out Llama 3 models ‘very soon’

Meta isn’t ready to unveil the entirety of its Llama 3 large language model (LLM) just yet, but that isn’t stopping the company from teasing some basic versions “very soon”, the company confirmed on Tuesday.

Speaking at an event on Tuesday, Meta president of global affairs Nick Clegg said that the company will begin rolling out some of its Llama 3 models sometime this month.

“Within the next month, actually less, hopefully in a very short period of time, we hope to start rolling out our new suite of next-generation foundation models, Llama 3,” Clegg said, according to TechCrunch. “There will be a number of different models with different capabilities, different versatilities [released] during the course of this year, starting really very soon.”

Clegg’s comment confirms a report earlier on Tuesday from The Information, which cited people who claim to have knowledge of Meta’s plans and said the company could launch two small, non-multimodal versions of Llama 3 next week before releasing the biggest Llama 3 model, which would be multimodal, as part of a broader Llama 3 release this summer. Llama 3 has long been expected to deliver multimodal support, letting users input text as well as images to return responses.  

Also: I tested Meta’s Code Llama with 3 AI coding challenges that ChatGPT aced – and it wasn’t good

Llama 3 is Meta’s answer to OpenAI’s GPT-4, Anthropic’s Claude 3, Google’s Gemini, and other LLMs. Meta launched Llama 2 in July 2023 as an open-source platform, letting anyone use it as they see fit. Llama 2 was largely successful in helping Meta get a spot at the AI for businesses table, but the company still trails OpenAI and others for market leadership.

Meta has been tight-lipped regarding its broader Llama 3 plans, though most market analysts expect the company will undoubtedly ship Llama 3 as an open-source model. 

To say the stakes are high for Llama 3 would be an understatement. As companies around the globe race to adopt AI, the companies that are developing the LLMs are similarly trying to stay a step ahead — or catch up. Meta finds itself behind some of its competitors and absent a major leap ahead in 2024, runs the risk of remaining one of the companies trailing OpenAI.

Meta isn’t leaving anything to chance. The company has spent billions of dollars on AI investments and has acquired hundreds of thousands of Nvidia H100 GPUs. In other words, Meta is ready for a fight with other tech heavyweights.

Also: Meta’s ‘pruning’ of Llama 2 model shows path to slimmer AI

It’s unclear why Meta would want to tease Llama 3 next week. It’s possible the company wants to showcase some of its better innovations to whet the appetite for those who are waiting to determine which model they want to use later this year. It’s also possible that Meta wants to make clear that what it has in store will be something to watch — and wait for. Or maybe, Meta doesn’t want to look like it’s already lost the race.

Whatever the case, we likely won’t have long to wait. After Meta launches Llama 3 updates, the company is expected to launch the full model globally sometime this summer.

Update at 11:52 a.m. ET on April 10 to include Meta’s confirmation that Llama 3 models will be launching soon.