Meta has launched Muse Spark Open, a new open-weight AI model that runs locally on Apple silicon, in collaboration with Apple. The model is designed for on-device execution on Macs rather than requiring cloud inference. The source material does not specify model size, licensing terms, or performance benchmarks.
Meta has introduced Muse Spark Open, an open-weight model with 816 billion parameters that handles text, images, video and audio with context windows up to one million tokens. According to the report, it performs roughly on par with the most capable open-weight Chinese models on benchmarks and is optimized to run locally on Apple's M5 Ultra. The release positions Meta against Qwen, GLM and Kimi in the open-weight race.
Meta's coding-optimized open-weight model Muse Spark Open is now available in Amazon SageMaker JumpStart, with BF16 and NVFP4 quantized variants. The MoE model has 816B total parameters (42B active), a 1M-token context window, and accepts text, image, document, audio, and audio-video input. It deploys to SageMaker AI endpoints in the user's own AWS account via Studio or the Python SDK, with VPC, KMS encryption, and network isolation options.
NVIDIA released Nemotron 3.5 Super VL, an open 120-billion-parameter vision-language model aimed at multimodal agentic tasks such as tool use, instruction following and document understanding, along with open weights, datasets and training recipes. It also introduced two edge models, Nemotron 3.5 Bolt 74B-A9B 24GB for local agentic workloads and Bolt E4B 4GB for conversational and physical agents. Super VL and Bolt 74B-A9B are available now, with Bolt E4B due Oct. 23. NVIDIA says several organiz
Why it matters: It gives readers concrete details on three newly released open models spanning data center to edge, including deployment options, availability and enterprise customization efforts.
