Every idle GPU, answering as one.
Covey pools spare GPUs from labs and data centers into a single inference endpoint. You send one request; the nearest free card answers it.
- GPUs online
- 41,207
- P50 latency
- 38 ms
- Per 1M tokens
- $0.18
Covey pools spare GPUs from labs and data centers into a single inference endpoint. You send one request; the nearest free card answers it.