“Meta on Wednesday debuted Muse Spark, a homegrown AI model it says significantly narrows the performance gap with models from OpenAI, Anthropic and others.”
“Why it matters: The model — code-named Avocado and built over the past nine months by a team led by Alexandr Wang — is a major upgrade over its Llama 4 models, Meta tells Axios.”
Key here is not how good Muse Spark is vs the big guys’ model. But more is it good enough for Meta’s 3.5+ billion daily users.
“Driving the news: Muse Spark will power queries in the Meta AI app and Meta.ai website immediately, with plans to expand across Facebook, Instagram and WhatsApp.”
“The model accepts voice, text and image inputs, but produces text-only output.”
“As Axios first reported, Meta plans to release a version of Muse Spark under an open-source license.”
So far, so good. It has a beat, and you can dance to it. But there’s a bit more:
“Zoom in: The model uses a fast mode for casual queries and several reasoning modes.”
“A “shopping mode” highlights how Meta hopes to differentiate itself. It combines large language models with data on user interests and behavior.”
“Over time, the model will also power “features that cite recommendations and content people share across Instagram, Facebook, and Threads,” Meta said in a blog post.”
And of course, Meta will integrate it well into its world class AI driven tools for advertisers. Even Google may not be as good as Meta in this department.
Muse Spark may not have all the bells and whistles, but it should be able to handle the basics for billions. At efficient costs.
“Zoom out: A Meta executive told Axios that Muse Spark doesn’t mark a new state of the art, but is competitive with the latest models from leading labs at certain tasks, including multimodal understanding and processing health information.”
“In other areas, including coding, the company acknowledges that there is a gap between Muse Spark and the models already available.”
“All flavors of the model are free to use, though Meta may impose rate limits.”
And then of course there’s the privacy thing. Which is not really a thing at Meta.
“Yes, but: Consumers should be aware that Meta’s privacy policy sets few limits on how the company can use any data shared with its AI system.”
“The big picture: Meta sees the release of Muse Spark as just one step toward its broader vision of personal superintelligence.”
And of course, the world is going to be focused on Meta’s new model vs the existing leaders. Who’re ramping up their own newer, bigger LLM AI models. AI Agents and all.
“However, Meta’s rivals are also leaping forward. Anthropic this week detailed Mythos, a model it says is so powerful that its initial release is being limited to a handful of tech companies to focus on cybersecurity defense.”
“OpenAI is also putting the finishing touches on a new model, code-named Spud, that is also believed to represent a significant leap forward.”
But the real takeaway here should be that Muse Spark should be good enough for Meta’s consumer Meta AI product. It’s distribution is far better than OpenAI and Anthropic, the two big dogs in the race.
Good morning, This week on Trends with Friends, we’ve got a killer lineup: Mike Dudas joins us to break down why the crypto market is roaring back to...
Today’s topic brings to mind the ‘Spy vs Spy’ cartoons of Mad Magazine fame from a few decades ago. As I’ve outlined earlier, the next wave of innovation...
