Meta's Muse Spark Open (816B coding MoE) lands on SageMaker with NVFP4 variant
AWS ML Blog · rss · 2026-10-16
AWS announced that Meta's coding-optimized open model Muse Spark Open (816B total/42B active MoE, 1M-token context) is now available in SageMaker JumpStart, in BF16 and NVFP4 variants.
- NVFP4 4-bit weights need roughly a quarter of BF16 memory, running on ml.p5e.48xlarge; BF16 requires ml.p6-b300.48xlarge
- Built for long-horizon agentic coding: multi-file refactors, long debugging, web browsing, subagent coordination, and computer use
- Endpoints run in your own AWS account with VPC, KMS encryption, and network isolation
The post includes a full deployment walkthrough via SageMaker Studio and the Python SDK.
More from Infra
- Berkeley scholars: concentrated chip supply chain makes a hardwired frontier AI pause feasible — wfithian · 2026-10-16
- Microsoft trained MAI-Base-2 on 16k GB200 GPUs in just 38 days, with 21% MFU — scaling01 · 2026-10-16
- Robotics data tokenization landscape: Khala maps peaq, THOR, BitRobot and Axis — 0xSammy · 2026-10-16
- Signal65 benchmarks: Intel Xeon 6 runs agent workloads 11% faster for less money — ryanshrout · 2026-10-16
- Microsoft's MAI-Base-2: 2.8T-param MoE pretrained on 16k GB200s in just 38 days — scaling01 · 2026-10-16
- Cloudflare and OpenAI pilot: real-time network signals bring unseen pages into ChatGPT search — irvinebroque · 2026-10-16