# RuntimeWire — Meta says Prometheus delivered 85% goodput while training Muse Wonder

- Company: RuntimeWire (runtimewire.com)
- Announced: 2026-10-20T01:18:06+00:00
- Category: infrastructure-release
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://runtimewire.com/article/meta-prometheus-muse-wonder-training-reliability
- Record: https://forck.live/items/24965-meta-says-prometheus-delivered-85-goodput-while-training-muse-wonder
- Subject: Reporting
- Models affected: Muse Wonder

Meta says its Prometheus supercluster in New Albany, Ohio, delivered more than 85% goodput during Muse Wonder pretraining and more than 90% during reinforcement learning. AI at Meta said in an October 19th post on X that the first phase of Prometheus trained Muse models, including Muse Wonder. The self-reported figures put reliability alongside GPU scale in Meta's frontier-model strategy and show the operational bottleneck it is trying to address as it builds toward a roughly 1 million-GPU cluster.

## Evidence

Verbatim from https://runtimewire.com/article/meta-prometheus-muse-wonder-training-reliability:

> Meta says its Prometheus supercluster in New Albany, Ohio, delivered more than 85% goodput during Muse Wonder pretraining and more than 90% during reinforcement learning .

---

Record: https://forck.live/items/24965-meta-says-prometheus-delivered-85-goodput-while-training-muse-wonder
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md