Introducing Muse Glimmer
- Source
- simonwillison.net
- Date
Context
Simon Willison's local trial shows Meta's Muse Glimmer being used to investigate a codebase. Running through LM Studio and a coding- plugin, the model made tool calls to explore a fresh copy of Datasette in response to the question of how its authentication works. Willison's account links the full transcript of those calls. The example demonstrates a local model participating in repository exploration; it does not establish the answer's correctness or the model's success rate across coding tasks.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- open weights — A model whose trained parameters are published for anyone to download and run — unlike API-only models you can access but never possess.
- tool use — A model's ability to call external functions — run code, search the web, edit files — instead of only generating text.
- SWE-bench — The standard benchmark for AI coding agents: real GitHub issues from real repositories, scored by whether the agent's patch passes the project's own tests.
“Meta are back in the open weights game! Muse Glimmer is a brand new 30B model under a clean Apache 2.0 license (a step up from the janky Llama licenses of old).”
Simon Willison
“Muse Glimmer achieves strong success rates on full-task benchmarks including DeepSearch QA, MCP-Atlas, 𝛕-Bench and SWE-Bench, which measure its ability to work within scaffolds, write and debug code, and resolve multi-turn requests from start to finish.”
Meta
“I really like this size of model, because if a machine has 32 GB of RAM or more (mine has 128GB) it leaves plenty of space for running other applications at the same time.”
Simon Willison
Checking sign-in…
Loading comments…