ElephantPool logo ElephantPool
Models Rete How it works Help Blog Sign in
Sign in

The network, right now

Which GPUs are serving which models, how fast they measured, and what people asked for that nobody could serve. Anonymised, and the same numbers whichever side of the market you are on.

The network view is unreachable right now. The API itself is unaffected.

From here

Model catalogue · Connect a GPU · Choosing what your GPU serves · What makes a GPU fast

ElephantPool logo ElephantPool

Open-weight AI inference served by community GPUs across Europe. Builders pay less; hardware owners get paid for every token they serve.

Product

Models Rete How it works Run a GPU Sign in

Resources

Knowledge base Whitepaper Technical spec mahout source Contact

Blog

one model, every card: how a 27b runs on 8 gbwe stopped projecting your earnings and started measuring themprices should not need an engineerour website and our app were two products. now they are one. All posts →

Legal

Conditions d'utilisation Mentions légales Privacy policy
© 2026 ElephantPool · De Buck Technologies SRL, Brussels, Belgium · BE 0787.873.392