{"version":"1","article":{"id":"muse-glimmer-on-device-agent-model-2026","slug":"muse-glimmer-on-device-agent-model-2026","language":"en","url":"https://haberpaper.com/en/news/muse-glimmer-on-device-agent-model-2026","api_url":"https://haberpaper.com/api/v1/articles/muse-glimmer-on-device-agent-model-2026?lang=en","title":"Muse Glimmer 30B: An On-Device Agentic Model","summary":"A research preprint introduces Muse Glimmer, an open-source 30-billion-parameter causal language model with a perception encoder for agent tasks on consumer hardware. The authors describe text and image input, multi-step reasoning, tool use and failure recovery.","category":"business","region":"world","topics":["ai","semiconductors"],"published_at":"2026-10-20T19:09:32+03:00","modified_at":"2026-10-20T19:09:32+03:00","publication_basis":"publication-record","event_period":"2026-10-17","event_period_basis":"event-or-reference-period","sources":[{"name":"arXiv","url":"https://arxiv.org/abs/2610.24628v1","publishedAt":"2026-10-17","checkedAt":"2026-10-20"}],"translations":[{"language":"tr","url":"https://haberpaper.com/tr/news/muse-glimmer-on-device-agent-model-2026","api_url":"https://haberpaper.com/api/v1/articles/muse-glimmer-on-device-agent-model-2026?lang=tr"},{"language":"en","url":"https://haberpaper.com/en/news/muse-glimmer-on-device-agent-model-2026","api_url":"https://haberpaper.com/api/v1/articles/muse-glimmer-on-device-agent-model-2026?lang=en"},{"language":"ar","url":"https://haberpaper.com/ar/news/muse-glimmer-on-device-agent-model-2026","api_url":"https://haberpaper.com/api/v1/articles/muse-glimmer-on-device-agent-model-2026?lang=ar"},{"language":"fr","url":"https://haberpaper.com/fr/news/muse-glimmer-on-device-agent-model-2026","api_url":"https://haberpaper.com/api/v1/articles/muse-glimmer-on-device-agent-model-2026?lang=fr"},{"language":"es","url":"https://haberpaper.com/es/news/muse-glimmer-on-device-agent-model-2026","api_url":"https://haberpaper.com/api/v1/articles/muse-glimmer-on-device-agent-model-2026?lang=es"}],"data_types":[],"data":null,"data_compilations":[],"data_downloads":[],"content":{"paragraphs":["A research preprint introduces Muse Glimmer, an open-source 30-billion-parameter causal language model with a perception encoder for agent tasks on consumer hardware. The authors describe text and image input, multi-step reasoning, tool use and failure recovery.","The stated aim is local operation without cloud infrastructure. The paper is the primary source for the model description; it does not by itself establish independent benchmark performance or broad device compatibility."],"key_points":[],"sections":[]},"text":"A research preprint introduces Muse Glimmer, an open-source 30-billion-parameter causal language model with a perception encoder for agent tasks on consumer hardware. The authors describe text and image input, multi-step reasoning, tool use and failure recovery.\n\nA research preprint introduces Muse Glimmer, an open-source 30-billion-parameter causal language model with a perception encoder for agent tasks on consumer hardware. The authors describe text and image input, multi-step reasoning, tool use and failure recovery.\n\nThe stated aim is local operation without cloud infrastructure. The paper is the primary source for the model description; it does not by itself establish independent benchmark performance or broad device compatibility.","corrections":[],"editorial_illustration":{"kind":"symbolic-editorial-illustration","documentary":false,"scope":"section","asset_id":"business","url":"https://haberpaper.com/assets/illustrations/business-1536.webp","width":1536,"height":1024,"alt":"Symbolic industrial workshop with machinery, a warehouse and a crane.","language":"en"},"section_illustration":{"kind":"symbolic-editorial-illustration","documentary":false,"scope":"section","asset_id":"business","url":"https://haberpaper.com/assets/illustrations/business-1536.webp","width":1536,"height":1024,"alt":"Symbolic industrial workshop with machinery, a warehouse and a crane.","language":"en"}}}