Meta releases Muse Spark Open: 816B open-weight model that runs on two M5 Ultra Mac Studios
op7418 · x · 2026-10-16
Meta released Muse Spark Open, billed as the most capable US open-weight model to date. Key points:
- Specs: 816B total / 42B active parameters, 1M-token context, natively multimodal (text, image, video, audio).
- Local deployment: first in its class to run on desktops—optimized with Apple's MLX framework, using the M5 Ultra's unified memory (up to 512GB per machine) across two Mac Studios.
- Positioning: built for coding and agentic work—long-horizon tasks, complex software engineering, multi-step reasoning; reported benchmarks are roughly on par with Kimi K3 and GLM-5.3.
- Availability: live on Baseten, Fireworks, Groq, OpenRouter, Together AI and others; weights on Hugging Face, SageMaker, Google Cloud; supports SGLang, vLLM, TensorRT-LLM, llama.cpp with AMD/Apple/Intel/NVIDIA optimizations.
The accompanying eval methodology doc details inference parameters used when benchmarking against Inkling, Mistral Large 4, Nemotron 3 Ultra, Qwen and others.
Related event: Meta Open-Sources Muse Spark Open with 816B Parameters and 1M Context(39 posts)→
More from Models
- Muse Spark Open runs locally on a Mac M3 Ultra via llama.cpp: frontier model on desktop — pcuenq · 2026-10-16
- OpenAI ships GPT-Live-1.1: new voices in six languages, fewer confirmations — OpenAI · 2026-10-16
- Muse Spark model appears on LMArena, testable in Battle and Agent modes — arena · 2026-10-16
- Running a multi-model UI design arena: Paseo orchestrates four design drafts per model — vista8 · 2026-10-16
- Ilya Sutskever questions Meta license clause taxing MaaS providers above $20M revenue — ilyasu · 2026-10-16
- Muse-Spark open-weights drop: 816B-A42B MoE model gets 359GB MLX 3-bit build — ns123abc · 2026-10-16