FULL STORY
CrowdGPT: Training an Open LLM with Crowdsourced GPUs
A developer launched CrowdGPT on Reddit to train a 1B-parameter open LLM using community GPUs with no datacenter. Within a week, 85 contributors fed 13 billion tokens into Crowd-v1.
2026-10-25 ~ 2026-11-08 · 2 episodes · 5 posts
Episode 1 · CrowdGPT aims to crowdtrain an open LLM without data centers (2026-10-25, 2 posts)
Developer Vxtzq1 launched CrowdGPT, a community project to train a 1B-parameter open LLM using volunteers' idle GPUs, with about 690M tokens trained so far and no data center required.
- CrowdGPT: a crowd-trained, datacenterless 1B open-source LLM, 6.9B tokens done — Vxtzq1 · 2026-10-25
- CrowdGPT Aims to Train a 1B-Parameter LLM Collaboratively Without Datacenters — Vxtzq1 · 2026-10-26
Episode 2 · CrowdGPT Crowdsources GPU Power to Train First Datacenter-Free 1B Open LLM (2026-11-08, 3 posts)
Open-source nonprofit project CrowdGPT is training Crowd-v1, a 1B-parameter transformer built entirely through crowdsourced GPU contributions with no datacenter and zero cost. In about a week, 85 contributors have already supplied 13 billion tokens, with participants gaining ownership of the resulting model.
- CrowdGPT trains a 1B collaborative LLM with zero dollars: 85 contributors, 13B tokens in a week — Vxtzq1 · 2026-11-08
- CrowdGPT Wants to Train an LLM Collaboratively—No Datacenter, Fully Open Data and Weights — Vxtzq1 · 2026-11-08
- CrowdGPT Wants to Train an Open-Source LLM on Crowdsourced Home GPUs — Vxtzq1 · 2026-11-08