Meta has shifted to a new flagship AI model called Muse Spark and sent the Llama "herd" of models to the farm upstate. It's a dramatic but understandable shift in AI strategy from the company that was once pioneering open models.Â
Muse Spark is a multimodal reasoning model, and unlike the Llama models, it's a closed-weight proprietary model. You won't be able to download it from Hugging Face or run it on your own machine for any reason.
There are several other AI products also under the Muse banner: Muse Glimmer is an open-weight model for running agents locally, Muse Code is Meta's take on Claude Code and Codex, and Muse (with no other modifier) is its personal AI agent app. The company has also announced Muse Charm, a voice-based physical AI device.
As a fan of open models, I'm sad to see Meta pivot away from Llama, but it's understandable given the rising competition from Chinese open models from DeepSeek and Moonshot AI, as well as the need to stay competitive with the other American AI labs. So, let's look at how Muse Spark, Muse Glimmer, and the rest of Meta's AI tools stack up.
Table of contents:
What is Muse Spark?
Muse Spark is the leading AI model developed by Meta Superintelligence Labs. Like other frontier models, Muse Spark is a multimodal reasoning model that supports tool use and agentic behaviors. In other words, it can analyze visuals as well as text, decide whether to generate an image in response to a question or search the web, and have sub-agents handle tasks. Right now, we're at Muse Spark 1.3.
Muse Spark is legitimately a frontier model. It's consistently ranked in the top handful of models on Artificial Analysis, behind only the latest models from OpenAI and Anthropic. (Though it's worth noting the Muse Spark version currently available to the general public and not select partners is a few points less powerful.)
Muse Spark is meant to be the first in a family of models, though so far, we've mostly had Spark point releases. While details are still a bit scarce about other models coming down the pipeline, when it launched, Muse Spark was apparently the "first step on [Meta's] scaling ladder and the first product of a ground-up overhaul" of its AI program. We're still waiting on more models—there's only been the open-weights Muse Glimmer and the image-generating Muse Image—but it's safe to say the AI program has been overhauled with Muse, Muse Code, and Muse Charm all launching over the last few months.
What is Muse Glimmer?
Muse Glimmer is an open-weight model from Meta Superintelligence Labs. It has 30 billion parameters and is designed to run AI agents locally. A Mac or PC with a powerful GPU is all that you need to get started: you can download it from Hugging Face right now. Â
Muse Glimmer is far from the most powerful AI model available, but then, it's not meant to be. The whole point is that it is small enough to run on consumer hardware. When compared to similar models from the Gemma and Qwen families, for example, it performs well.Â
I'm glad that Meta hasn't totally given up on open models, and I'm curious to see how Glimmer (and its successors) hold up over the next while.
What is Meta Muse?
Muse (just Muse) is an agentic personal assistant app that uses the Muse Spark model. Mark Zuckerberg has talked about building "personal superintelligence," and this seems like its first real step in that direction.Â
From an architecture standpoint, Muse is fascinating. While you talk to it in a chat window, every user actually gets their own virtual Linux cloud computer called Muse Secure VM. Your agent operates on it, but you can also step in and use it too. It's configured so that the Muse agent can use your passwords and login credentials without seeing them (though there are still security concerns). It can browse the web, make purchases, and connect with third-party apps and tools. There's a second agent, Sentinel, that's meant to watch over every online action Muse takes to make sure it's safe.
Since Muse is running on its own computer, it can keep working even when your phone is off. Use the Muse app to tell it to do something—like sell things on Facebook Marketplace or find a restaurant—and it will keep working until it's done. It'll check with you before certain actions like sending an email or making a purchase, but you don't need to constantly watch it.Â
As you can imagine, with features like that, Muse has huge potential to cause chaos. Amazon has already blocked Muse from its online store. One user says his Muse agent cheaply sold items on Facebook Marketplace then gave out his home address so that the buyer could collect it. A journalist for Inc. says Muse went through his iMessage archives without permission. Both cases seem to be more complicated than just the agent going rogue, but I suspect lots more stories like these if Muse gets more popular—and it currently tops the iOS and Android download charts.Â
What is Muse Code?
Agentic engineering has proven to be the killer AI feature for OpenAI and Anthropic, so Meta obviously wants in on the action. That's why we have Muse Code. It's a terminal-based coding agent available on macOS, Linux, and Windows that uses Muse Spark. And in terms of coding, Muse Spark is pretty respectable. Like in most benchmarks, it's not quite as strong as the top models, but it hangs comfortably with the likes of Grok, DeepSeek, Kimi, and cheaper models from OpenAI and Anthropic.
One big thing it has going for it is the pricing: It has $5/month, $15/month, and $50/month tiers. Claude, on the other hand, has $20/month, $100/month, and $200/month tiers. While things obviously aren't completely comparable, I suspect there are a lot of people on $20/month Claude plans hitting their limits but unable to justify spending $100/month for whom a $50/month Muse plan with 200+ prompts every five hours might make a lot of sense.
Muse Spark is also available as an API. Meta has two versions of its models:
Muse Spark 1.3 costs $1.25 per million input tokens and $4.25 per million output tokens.Â
Muse Spark 1.3 Contributor costs $0.10 per million input tokens and $0.20 per million output tokens. It's the same model at more than a 90% percent discount if you let Meta train its models on your prompts and outputs.Â
It's the only frontier model where you can literally pay your token bills with your data.Â
How to try Muse Spark
Meta AI, the AI assistant built into Facebook, Messenger, Instagram, and WhatsApp, now uses Muse Spark. The best place to check it out is the dedicated web app.
If you're feeling adventurous, you can also check out Muse Spark through the Muse app or Muse Code app. For developers, there's an API available.
You can also add Meta's models to your workflows with Zapier. Build automations wherever you're already working, whether that's Muse or another AI assistant, and then let Muse Spark be the model doing the thinking inside it. Once it's built, it runs on Zapier on a schedule or trigger, and can take action across 9,000+ apps. You decide what the AI can and can't touch, and Zapier manages and governs your automations.
What happened to Llama?
In short, Llama was outcompeted by Chinese research labs like DeepSeek and Qwen. By the time Llama 4 shipped in April 2025, it was already trailing DeepSeek V3 and other top open models. The announced Behemoth and reasoning models that were meant to be competitive never shipped. Also, Llama 4's benchmark performance was overstated as specifically tuned versions were used to inflate scores.Â
All told, Llama 4 never had the impact that Llama 3 did, and the AI landscape moves incredibly quickly.Â
By June 2025, Meta was reorganizing. They formed Meta Superintelligence Labs, hired talent from OpenAI, Anthropic, and Google with wild compensation packages, had a new Chief AI Officer, went through a round of layoffs, and otherwise fully restructured their approach to AI. It's clear that, at this point, the company had decided that open source models were not its future.
Because they're open, you can still download the various Llama models. They're just out of date. Muse Glimmer is Meta's latest open-weight model, and it competes directly with Google Gemma and OpenAI's oss models.
What's next for Meta?
Meta is still going through a massive AI shift. Muse Spark, Muse Glimmer, Muse, and Muse Code are just the earliest results from Meta Superintelligence Labs and the pivot to Personal Superintelligence, but it's safe to assume that there's a lot more coming down the pipeline. It's unclear how many more models there will be in the family, whether they'll be more or less powerful, and whether Muse Charm will finally bring voice-controlled AI devices to the masses.
Personally, it looks to me like Meta has lost a bit of time in this transition period. Over the past year, AI agents and particularly AI coding agents have become increasingly important. They're what justify enterprise and high-value AI subscriptions. While Meta is still trying to catch up to Anthropic and OpenAI, I wouldn't be surprised if it ends up doubling down on trying to be the personal AI embedded in the apps you use—not another developer tool.Â
Either way, Meta is no longer a pioneering open source AI lab. It's another large tech company duking it out with its proprietary models.
Related reading:
This article was originally published in August 2023. The most recent update was in September 2026.
