gen‑ai.news
← Back
Multimodal

Meta is back with Muse Glimmer: local, agentic, multimodal, and open source

Meta has released Muse Glimmer, an open-source multimodal model that can run entirely on local hardware while supporting agentic workflows - meaning it can plan and execute sequences of tasks rather than simply responding to single prompts. The release is hosted on Hugging Face, continuing a pattern of Meta publishing model weights and code openly for researchers and developers to use and build on.

Glimmer's multimodal design means it can work across both visual and textual inputs, handling tasks that require understanding images, generating new visual content, and reasoning across modalities in a single system. This kind of integration is increasingly important for agentic applications, where a model needs to perceive its environment, decide on actions, and carry them out - all in sequence.

The local-first approach is a notable aspect of this release. Running a capable multimodal model without sending data to a remote server has practical implications for privacy, latency, and offline use. As capable open models continue to shrink in resource requirements, local deployment becomes a more realistic option for developers working outside of cloud infrastructure.

With Glimmer, Meta continues to position itself as a significant contributor to the open generative AI ecosystem. The combination of agentic behavior, multimodal input and output, and open weights under a permissive release makes this a model worth watching for anyone building tools that need to see, reason, and act across complex tasks.

Enjoy this story? Get the next one in your inbox.

Twice a week: the most important stories in generative image and video AI, distilled into a 2-minute read.

Free. Unsubscribe any time. No spam, ever.

Your next read

Twitch streamers can now opt out from training Amazon’s AI
Multimodal

Twitch streamers can now opt out from training Amazon’s AI

Twitch has introduced an opt-out setting that lets streamers prevent their content - including streams, VODs, clips, chat logs, and channel images - from being used to train Amazon's generative AI models. The control applies to future training only and covers AI systems designed to generate or synthesize text, audio, images, or video. Non-generative AI features such as captions and safety tools are unaffected by the setting.

Claude will apply invisible watermarks to AI text and images
Multimodal

Claude will apply invisible watermarks to AI text and images

Anthropic has announced plans to embed invisible watermarks into text and images produced by Claude, making it easier for platforms and users to identify AI-generated content. The move is tied to European regulatory requirements around AI transparency. The changes are not yet live, but represent a formal commitment from the company.

Adobe Announces It Is Further Integrating Into ChatGPT With New Unified Plugin
Multimodal

Adobe Announces It Is Further Integrating Into ChatGPT With New Unified Plugin

Adobe and OpenAI have deepened their existing partnership with a unified plugin that brings more than 70 Adobe tools directly into ChatGPT. The integration spans apps including Photoshop, Firefly, Premiere, Illustrator, and Acrobat, letting users describe a creative goal and have the AI orchestrate the appropriate tools to complete it. The plugin is available now for ChatGPT users, with guest access offered and fuller capabilities unlocked via an Adobe account sign-in.