Meta has returned to open models with the release of Muse Glimmer, the first open model from its rebuilt AI group, Meta Superintelligence Labs, following the flopped Llama 4 and the departure of chief scientist Yann LeCun. Meta compares Glimmer with Google's Gemma4-31B and Alibaba's Qwen3.6-27B, claiming Glimmer wins most benchmarks, especially on agent tasks like tool use, web search, and long-context work, though Qwen is better at driving a computer desktop and terminal tasks, and all three are even on multimodal tasks. Meta gathered most of the comparison data itself and admits its test setup isn't tuned for rival models. At full precision, the model would need more than 55 GB of memory, but Meta squeezes the weights to about 4 bits, pushing the model under 20 GB, fitting into consumer graphics cards and MacBooks. A small helper model speeds up text output by up to 3.1x. Glimmer was trained by distilling Meta's larger Muse Spark model. Alongside the release, Zuckerberg published an essay titled "The Future is for Everyone," arguing superintelligence should be spread widely and bluntly defending distilling other companies' models, writing that the principle worth protecting is "that you can learn from anything you can observe." This lands amid an ongoing fight where OpenAI and Anthropic have accused Chinese labs of using their models without permission, and Anthropic CEO