Muse Spark Open (AI model): news timeline & CVEs

[AINews] Agents disprove Polysemanticity; greatly improve Interpretability & Alignment

Meta released Muse Spark Open, an 816B MoE that leads US open-weight models on Artificial Analysis.

Meta released Muse Spark Open, an 816B-total / 42B-active mixture-of-experts model with a 1M-token context and text, image, video, and audio input. Artificial Analysis scores it 39, first among US open-weight models but behind GLM-5.3 (45) and Kimi K3 (44); launch pricing is $1.25/$4.25 per million input/output tokens. The same roundup covers a Goodfire interpretability essay and launches of NVIDIA Nemotron 3.5 Super VL and Microsoft’s roughly 3T-parameter MAI-Base-2.

Latent SpaceOct 16, 04:16 UTC · 21h agoModel release 7 sources2· 1 read

Meta Releases Muse Spark Open: 816B Open-Weight MoE With 1M Context

Meta released Muse Spark Open, an 816B open-weight MoE with a 1M-token context.

Meta released Muse Spark Open, an 816-billion-parameter mixture-of-experts model with open weights and a one-million-token context window. The captured page is almost entirely cookie-consent text, so benchmark scores, license terms, and availability details are not stated.

MarkTechPostupdated Oct 16, 04:16 UTC · 21h agofirst Oct 16, 02:26 UTC · 23h agoModel release 7 sources2· 1 read

Meta launches an open-weight AI model that runs on Macs

Meta released Muse Spark Open, an 816-billion-parameter open-weight model built to run locally on Macs.

Meta launched Muse Spark Open, calling it the most capable open-weight model in the United States. The natively multimodal model has 816 billion parameters and a 1 million-token context window, and is built for coding and agentic tasks. Meta says local use requires at least 1 terabyte of RAM and works best on a Mac Studio with an M5 Ultra chip. It demonstrated the model with Muse Code and is making the weights available through API platforms.

TechCrunch · AIupdated Oct 16, 04:16 UTC · 21h agofirst Oct 15, 20:04 UTC · 1d agoModel release 7 sources3· 1 read

Muse Spark Open is Meta's answer to Qwen, GLM, and Kimi in the open-weight race

Meta released Muse Spark Open, an 816-billion-parameter open-weight model for coding, agents, and multimodal tasks.

Meta released Muse Spark Open, an open-weight model it calls the most capable US open-weight model, aimed at coding and agentic work. It has 816 billion parameters with 42 billion active, up to one million tokens of context, and text, image, video, and audio support. Meta says it leads Qwen 3.8, GLM 5.3, and Kimi K3 on agentic, knowledge, reasoning, and multimodal benchmarks but trails some rivals on coding and remains well behind paid Muse Wonder Full models. Weights are on Hugging Face under a free license that requires a "Built with Muse Spark Open" label; large API providers need a separate agreement.

The Decoderupdated Oct 16, 04:16 UTC · 21h agofirst Oct 15, 18:34 UTC · 1d agoModel release 7 sources2· 2 reads

Meta and Apple team up to make Muse AI models run better on Macs

Meta released Muse Spark Open, an 816B open-weight model tuned with Apple to run on Mac Studios.

Meta released Muse Spark Open, which it calls the most capable US open-weight model to date, and said it is the first model in its class to run on a desktop using two Apple M5 Ultra Mac Studios with MLX. The natively multimodal model has 816 billion total parameters and is described as strong at long-horizon tasks, software engineering, and multi-step reasoning. Meta and Apple publicly credited months of joint work, and Apple published a technical post on MLX, Muse Spark Open, and two other open models.

The Verge · AIupdated Oct 16, 04:16 UTC · 21h agofirst Oct 15, 18:15 UTC · 1d agoModel release 7 sources2· 2 reads

Muse Spark Open, the Most Capable US Open‑Weight Model to Date

Meta released Muse Spark Open, an 816B multimodal open-weight model with 42B active parameters and a 1M-token context.

Meta released weights for Muse Spark Open, a natively multimodal mixture-of-experts model with 816 billion total parameters, 42 billion active, and a one-million-token context window, built for coding and agentic work. Meta calls it the most capable US open-weight model and says it is the first in its class to run locally on two Apple M5 Ultra Mac Studios via MLX, using up to 512 GB of unified memory per machine. Weights are on Hugging Face, Amazon SageMaker, Google Cloud Model Garden, and Microsoft Foundry, with APIs at Baseten, Fireworks, Groq, Modal, Nebius, OpenRouter, and Together AI. The Muse Model License has limited commercial terms, and Meta says safety post-training met its Superintelligence Scaling Framework release thresholds.

Hacker News · AIupdated Oct 16, 04:16 UTC · 21h agofirst Oct 15, 16:58 UTC · 1d agoModel release 7 sources HN 57↑ · 4 comments

Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta

Meta introduced Muse Spark Open, an open-weight frontier multimodal mixture-of-experts model.

Hugging Face announced Muse Spark Open, described as an open-weight frontier multimodal mixture-of-experts model from Meta. The available notice does not include parameter count, benchmark scores, license terms, or other release specifics. It is a major open-weight model launch from a leading AI lab.

Hugging Face Blogupdated Oct 16, 04:16 UTC · 21h agofirst Oct 15, 00:00 UTC · 2d agoModel release 7 sources1

Muse Spark Open

Timeline7 articles

Appears with