Mistral Small 3 24B
The best latency per euro in the catalogue for European traffic.
22ms
p50 from the nearest warm region
364tok/s
Single stream, no batching
32K
Full window, no sliding truncation
$0.14
Per million tokens served
The best latency-per-euro in the catalogue for European traffic. Paris and Amsterdam both hold dedicated capacity for it.
Deploymentdep_d3d944
Parameters24B
Quantisationfp8
LicenceApache 2.0
Home regioncdg1
Extra replicas4
p50 latency24ms
p99 latency58ms
Requests / s2880
Figures are illustrative — this is a demonstration site.
Last 30 minutescdg1
cache hitqueue 0ms
errors 0