Muse Glimmer Unveiled
Meta has introduced Muse Glimmer, a new AI model designed for end-to-end task completion and reliable tool use. The model is licensed under Apache 2.0 and boasts strong performance on various benchmarks.

Meta has made a significant move in the AI landscape with the introduction of Muse Glimmer, a 30B model that operates under a clean Apache 2.0 license. This licensing choice marks an improvement over previous models, such as those in the Llama series. Muse Glimmer is optimized for tasks that require the model to work from start to finish, including complex operations like writing and debugging code. The model has demonstrated strong success rates on benchmarks like DeepSearch QA, MCP-Atlas, 𝛕-Bench, and SWE-Bench. Additionally, it is capable of handling a wide range of function calls, making it suitable for extended workflows that involve precise tool invocation. This development is noteworthy as it indicates Meta's continued investment in open and accessible AI technologies.
Read also

Meta Unveils Muse Ember: Ultra‑Cheap Model for Agent Execution
Meta launches Muse Ember, a lightweight model priced at 10 cents per million input tokens, positioning it as a cheaper alternative to leading competitors.

OpenAI Launches GPT‑6.2 Astra and GPT‑6.2 Sol Models
OpenAI adds two new GPT‑6.2 variants—Astra for complex projects and Sol for versatile daily use—both accessible through the API.

Nvidia Reports Strong Earnings Amid Jensen Huang Ouster Fallout
Nvidia posted solid earnings and reaffirmed growth prospects, even as it ends H200 chip sales in China following the Jensen Huang scandal.