Pluralis-8B decentralized training run: 98 consumer-GPU nodes at ~$11 per billion tokens
jon_durbin · x · 2026-09-11
jondurbin shares the Pluralis-8B decentralized training run: a dense 8.6B model (vs. comparable MoE setups) trained on 98 contributor nodes of consumer 4090/5090 GPUs over FineWeb-Edu 1.3T, at roughly $11 per billion tokens. A public Agora dashboard exposes telemetry, MFU, and per-node splits — a rare open experiment in consumer-GPU swarm training.
More from Infra
- DeepSpeed shows AutoEP/AutoTP/AutoSP parallelism gains, hosts office hours — PyTorch · 2026-10-27
- Cloudflare Durable Objects validated as top architecture for AI agent WebSocket servers — tristanbob · 2026-10-27
- Gary Marcus: With Q3 data added, Anthropic's compute-efficiency growth shows deceleration — GaryMarcus · 2026-10-27
- SpaceX locks in $40B GPU financing deal with Apollo, the largest ever committed — TechCrunch AI · 2026-10-27
- Inference is just Kubernetes with a multi-trillion-dollar market, and the communities overlap — matt_slotnick · 2026-10-27
- vLLM eyes Rust migration as inference engines race to rewrite the stack — AccBalanced · 2026-10-27