Fordel StudiosMeta is developing a new image and video model for a 2026 release, report says
What Happened
Meta aims to make the text-based model better at coding while also exploring new world models that understand visual information and can reason, plan, and act without needing to be trained on every possibility.
Our Take
Meta's chasing shadows again. Building a "world model" (fancy term for video prediction) while everyone's using Claude. Their coding improvement angle is a year late—OpenAI's been there already.
Real thing: if they nail visual reasoning at scale, robotics gets interesting. But 2026 release means 2027 reality, and everyone else will've shipped it by then.
Meta's actual problem isn't the model—it's trust. Enterprises don't want inference routing through them. They'd need to buy credibility, and they won't spend the money.
What To Do
If you need vision reasoning today, Claude's your move, not Meta's roadmap.
Cited By
React