Muse Spark Open launches: 816B-param MoE billed as the most capable US open-weight model
altryne · x · 2026-10-16
Scale AI's Alexandr Wang announced Muse Spark Open, billed as the most capable US open-weight model to date: 816B total / 42B active params, 1M-token context, natively multimodal, built for coding and agentic work. The shared benchmark table shows MCP-Atlas 89.2, Terminal-Bench 2.1 76.4, best-in-family HLE-Diamond 26.1 (no tools), ZeroBench 50.0, SimpleVQA 73.6; closed Muse Spark 1.3 still leads on most coding benchmarks.
Related event: Meta Open-Sources Muse Spark Open, an 816B-Parameter Flagship Model(46 posts)→
More from Models
- Are the fastest GPU kernels now written by AI? Kernels are hyper-verifiable, argues one dev — fleetwood___ · 2026-10-16
- Scale AI launches muse spark open: 816B-param MoE with 1M context, billed as top US open-weight model — jiayi_pirate · 2026-10-16
- NVIDIA HF listing for Qwen3.8-Flash-Next-FP4-Local leaks unannounced Qwen3.8 model — TheZachMueller · 2026-10-16
- Muse Spark Open, billed as the most capable US open-weight model, lands on Nebius day one — Arindam_1729 · 2026-10-16
- Model at capacity for a straight week drives frustrated dev back to Opus — joonasvirtanen · 2026-10-16
- Sentry CEO on expensing Claude subs: it's allowed, employers just want ZDR — zeeg · 2026-10-16