Model rankings
Throughput, latency, time to first token and price for every model on BlazeRail, measured on the requests each model actually served. A dash means the model served no requests in the window; it is not a zero and not an estimate.
Throughput, latency, time to first token and price for every model on BlazeRail, measured on the requests each model actually served. A dash means the model served no requests in the window; it is not a zero and not an estimate.