{"type":"post","id":"meta-s-vision-encoders-in-multimodal-models-6597ca","slug":"meta-s-vision-encoders-in-multimodal-models-6597ca","title":"Meta's Vision Encoders in Multimodal Models","description":"Meta researchers compared four ways to equip multimodal models with vision, training them under a controlled setup.","retrieval_nugget":"Meta researchers compared four ways to equip multimodal models with vision, training them under a controlled setup. Their results suggest that dedicated visual capacity matters more than separate encoder pretraining or a separate encoder boundary—a finding with implications for compute budgets and inference latency.","published_at":"2026-11-16","updated_at":"2026-11-16","record_date":"2026-11-16","date_kind":"published_at","topics":["multimodal"],"entities":["Meta","Unbundling Multimodal Vision Encoding","Ellis Brown","Chris Hoang","Shengyi Qian","Cambrian-7M","H200"],"related_patterns":[],"novelty_label":"new_fact","what_changed":"A new study finds that dedicated vision parameters matter more than separate pretraining or a pre-fusion encoder, with encoder-free models needing about three times the compute to match visual-question-answering performance.","prior_refs":["https://newruntime.com/posts/meta-ai-published-model-weights-on-hugging-face-2c8206/"],"source_url":"https://unbundling.github.io/","source_urls":["https://unbundling.github.io/"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:meta-s-vision-encoders-in-multimodal-models-6597ca","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/meta-s-vision-encoders-in-multimodal-models-6597ca/","markdown":"https://newruntime.com/posts/meta-s-vision-encoders-in-multimodal-models-6597ca.md","json":"https://newruntime.com/posts/meta-s-vision-encoders-in-multimodal-models-6597ca.json"},"markdown_url":"https://newruntime.com/posts/meta-s-vision-encoders-in-multimodal-models-6597ca.md","next_reads":[{"type":"topic","path":"/topics/multimodal/","url":"https://newruntime.com/topics/multimodal/","title":"Multimodal","reason":"Explore the multimodal topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/multimodal.md","json_url":"https://newruntime.com/topics/multimodal.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}
