Meta Releases Open-Weight Muse Glimmer, Reversing Course After Betting on Closed Muse Spark 1

Meta has released Muse Glimmer, an open-weight model, in a sharp reversal after betting on closed products, in particular the Muse Spark 1 model. According to AI Business, the change of course comes from pressure in the enterprise market: businesses demand the ability to deploy AI on their own infrastructure and not send data to the vendor's servers.

Why did Meta open the weights of Muse Glimmer?

Muse Glimmer is a direct response to requests from enterprise customers. Enterprise segments such as financial institutions, healthcare companies and government agencies cannot, as a matter of principle, send sensitive data to an outside provider's servers, even if that provider is Meta. Regulatory requirements, internal security policies and insurance risks make working through a cloud API physically impossible for part of the market.

The closed Muse Spark 1 gave Meta full control over the AI stack: from training to inference, from pricing to terms of use. But according to AI Business, this approach cut off an entire layer of potential enterprise customers who have no technical or legal ability to work through someone else's cloud API. The open weights of Muse Glimmer remove that barrier: the company gets the model and deploys it wherever it sees fit.

"This shift responds to enterprise demand for tighter data control and local infrastructure systems,"

AI Business writes in its coverage of Meta's reversal.

What does a business get with an open model?

Open weights are not just a "free download." An organization receives the model files themselves and deploys them wherever it sees fit: in its own data center, in a cloud of its choice, or on hybrid infrastructure. The data never leaves the company's secured perimeter, and traffic does not pass through Meta's servers.

For enterprise customers, this opens up capabilities that are fundamentally unavailable when working with a closed API:

• Full control over data — no traffic through Meta's infrastructure
• Fine-tuning for company tasks — fine-tuning on internal datasets without handing them to a provider
• Predictable scaling costs — no dependence on API pricing and terms that can change at any moment
• Independent security audit — an in-house team checks and validates the model on its own

How Meta's reversal fits the market

The decision to release Muse Glimmer fits a steady trend: major AI labs increasingly maintain parallel lines, with closed products that come with full support and a ready-made API for those who want to start fast, and open weights for those who need complete autonomy.

Closed models win where speed of adoption and minimal upfront operating costs matter and there are no specific restrictions on data transfer. Open models are indispensable in scenarios where regulatory requirements, data sensitivity or inference scale make dependence on an external API unacceptable. The split of enterprise demand into these two segments, apparently, became obvious enough for Meta to return to an open-weights strategy, even at the price of less control over the stack.

What this means

Muse Glimmer is an admission by Meta that closed models systematically fail to reach part of the enterprise market. The company is not abandoning Muse Spark 1, but it is opening a second front, in the place where a closed strategy simply does not work. For other AI labs, it is one more signal: in the enterprise, a presence in the open-weight segment is quickly shifting from a "competitive advantage" to a baseline market expectation.

*Meta is designated an extremist organization and banned in Russia.