@AIatMeta
Together with the AI community, we are pushing the boundaries of what’s possible through open science to create a more connected world.
Today we’re also previewing WildArtifactBench, an internal evaluation framework designed to assess agents on complex, real-world tasks across diverse deliverable formats. By using win rates and Elo scores from human and agentic preference judges rather than strict ground-truth rubrics, it expands task coverage across practical multimodal workflows. We’re releasing 10 tasks from WildArtifactBench as a step forward in our ability to measure the real practical utility delivered by multimodal agents: t.co/UIpdcw8JF2
We’re sharing some of our design principles for WildArtifactBench:
Muse Spark 1.2 supports a broad range of multimodal tasks, from turning visuals into working code to translating perception into physical action. It also brings robust audio-visual understanding to enable video-heavy workflows common in real-world enterprise use. Today, we’re sharing new evals and demos that illustrate the breadth of the model’s visual understanding and reasoning capabilities. Let’s start with a demo that shows how Muse Spark parses multimodal observations and calls tools to guide a robot to navigate in an unstructured environment to find a rubber duck. 🧵👇
Now let’s take a look at how Muse Spark 1.2 performs across visual reasoning, chart understanding, and knowledge-intensive tasks.
Introducing Muse Glimmer, an open-weight 30B-parameter model optimized for local, always-on agent workflows. Muse Glimmer delivers strong performance on key agentic use cases and benchmarks compared with leading models in its size category, and is designed to run entirely on consumer hardware like a Mac or PCs with performant GPUs. In keeping with our long tradition of sharing fundamental AI research, we’re releasing model weights under a permissive Apache 2.0 license. 🧵👇
For a local agent to be practical, generation latency must be low enough to maintain workflow continuity. To run Muse Glimmer on consumer hardware without degrading quality, we used quantization to shrink the language model to under 20GB and a lightweight DFlash drafter model to accelerate token generation. As a result, Muse Glimmer is fast enough for fluid conversation and real-time agent interaction, all running entirely on your device.
To understand whether we're making genuine progress on reasoning, we entered our AI models in five STEM Olympiad competitions. The results: 🏅 Asian Physics Olympiad (APhO): Perfect score, theory exam 🏅 International Physics Olympiad (IPhO): Perfect score, theory exam 🥇 International Mathematical Olympiad (IMO): Gold medal 🥇 International Chemistry Olympiad (IChO): Gold-medal-level performance 🥇 Romanian Masters of Mathematics (RMM): Gold-medal-level performance The types of problems in the Olympiad competitions are exceptionally hard, demanding deep chains of reasoning, creative insight, and flawless argumentation. To test pure reasoning capability, we disallowed all tool use, meaning no search, no coding, and no calculator. We have deep admiration for the contestants and committees behind these competitions, and are grateful for their support in enabling our participation.
Introducing Muse Code (beta), a terminal coding agent built for long-horizon software engineering, powered by our new Muse Spark 1.2 model. Muse Code plans, implements, and validates complex, multi-file changes across large repositories with persistent sub-agents that solve difficult problems faster, more accurately, and with less intervention. 🧵👇
Under the hood, Muse Code pairs a simple agent loop with persistent, async background agents that stay active throughout each session to execute next steps, avoid redundant info gathering, and reduce steering on complex tasks. An append-only local event log tracks every model call, tool run, approval, and edit, making the runtime replay-exact and restart-safe for long-running workflows.
To accelerate scientific discovery and support @ENERGY’s Genesis Mission, the @BerkeleyLab-led SYNAPS-I project is using SAM 3 and DINOv3 to automate image segmentation. By pairing DINOv3’s global semantic context and fine-grained spatial localization with SAM 3’s pixel-level boundary extraction, the researchers are able to compress 3D volume labeling from a month of manual effort to ~15 minutes. Learn more about their work: t.co/jBHRJPjFq5
We heard you and are happy to announce that Muse Spark 1.1 is now available on @OpenRouter for US-based developers. We look forward to seeing what the community builds. x.com/nuvolore/statu…
Get started: go.meta.me/e3b99d
To demonstrate Meta AI's advanced reasoning and multimodal capabilities, we submitted a model to participate in the Asian Physics Olympiad’s theoretical exam. We’re happy to share that our model achieved a perfect score of 30/30, tying with the top 3 student contestants. We appreciate the APhO committee for letting our model participate in the competition: t.co/dpyMyST2n4
We gave a few leaders early access to Muse Spark 1.1, here's what they had to say:
We’re excited to introduce Muse Spark 1.1, a significant upgrade from the first Muse Spark model we released earlier this year. Along with this release, we are launching a public preview of the new Meta Model API where developers can access Muse Spark 1.1. The model is also available now in "Thinking" mode in the Meta AI app and on t.co/wHkMPH82ZH. Learn more: t.co/zGcA3XaWpN
Muse Spark 1.1 is a multimodal reasoning model built for agentic tasks, with major gains in tool and computer use, coding, and multimodal understanding.
Straight from @finkd — Muse Spark 1.1 is live. x.com/finkd/status/2…
Alongside the release of Muse Image, we’re sharing an early preview of Muse Video. It offers competitive performance in prompt adherence, visual fidelity, and temporal consistency. We’re investing in areas with current performance gaps, such as audio-video synchronization and physically accurate fast motion.
Muse Image works as an agent rather than a direct prompt-to-image model: it invokes tools, self-refines, improves with scaled test-time compute, and pairs with Muse Spark for collaborative media generation. 🧵👇 x.com/AIatMeta/statu…
Muse Image writes and executes code to nail precise details like plots and QR codes, and teams up with Muse Spark to produce websites with embedded images and playable visual games. Explore this thread: meta.ai/share/c/z2nbtz…
Introducing Muse Image and Muse Video, the first media generation models developed by Meta Superintelligence Labs. Muse Image is our most advanced image generation model yet. It follows instructions faithfully, edits with precision, composes from multiple references, and draws on Instagram for social context. It also brings agentic tool use capabilities to image generation and integrates with Muse Spark. You can try Muse Image in the Meta AI app and web, as well as in Instagram Stories and WhatsApp – starting in limited countries with more locations on the way. Today we’re also previewing Muse Video, which is built upon the same pretraining base as Muse Image to deliver exceptional visual fidelity with native audio support. Learn more about both models: t.co/QtKDPDZP5v
We’re sharing the next major milestone in our non-invasive brain-to-text decoder research: Brain2Qwerty v2. Building on v1, which was published today in @Nature, Brain2Qwerty v2 is the highest-performing end-to-end pipeline capable of real-time sentence decoding from raw brain signals. It advances beyond character-level performance to decoding words and semantics, enabling accuracy for overall communication. We believe this research has the potential to make a real difference for the millions of people who suffer from brain lesions or disorders that prevent them from communicating. 🧵👇
We trained Brain2Qwerty v2 on ~22,000 sentences from 9 volunteers, each recorded for 10 hours wearing an MEG device while typing. By using end-to-end deep learning on raw brain signals from MEG devices and fine-tuning LLMs, the system effectively bridges the gap between noisy neural data and coherent language. The results are promising: - Avg word accuracy of 61% across participants - 78% word accuracy and 50%+ of sentences decoded with ≤ 1 word error for the top-performing participant - Performance scales log-linearly with data volume
Big congrats to our SAM 3D team for receiving a Best Paper Honorable Mention at #CVPR26! This prestigious recognition underscores their incredible work pushing the boundaries of computer vision. Read the paper here: arxiv.org/abs/2511.16624
Today we’re announcing an agreement with Amazon Web Services to bring tens of millions of AWS Graviton cores to our compute portfolio. This partnership marks an expansion of our diversified AI infrastructure and will help scale systems behind Meta AI and agentic experiences that
