Meta launches Muse Glimmer 30B, its first Superintelligence Labs model, on OpenRouter

Meta is releasing Muse Glimmer 30B, the inaugural model from its Superintelligence Labs, on OpenRouter starting 9 August 2026. The 30 billion-parameter model is dense, multimodal and supports text and images within a 131 k-token context window. Built on the Muse Spark architecture, it has been refined for reliable agent operation on consumer hardware such as an RTX 4090 or a MacBook Pro with large unified memory.

Muse Glimmer 30B is distributed under an Apache 2.0 license, allowing unrestricted commercial use, modification and redistribution. OpenRouter lists the usage cost at $0.30 per million input tokens and $1.20 per million output tokens.

In agent-oriented benchmarks, the model scored 75.5 on MCP Atlas and 51.2 on SWE-Bench Pro, metrics that evaluate multi-step planning, tool use, failure recovery and error correction without human intervention. The model also supports more than 100 languages and provides deep image understanding, enabling long-running autonomous workflows and code generation.

The launch marks the first public offering from Meta Superintelligence Labs, a unit created separately from Meta’s existing research groups FAIR and GenAI and focused on agent capabilities and general intelligence. If the agent-focused benchmarks remain community-aligned, Muse Glimmer could become the default open-source foundation for developers seeking a locally runnable alternative to closed APIs.