Qwen2.5 Coder 32B

Point an editor plugin at this one — fill-in-the-middle is a first-class field.

$argosy up --model qwen2-5-coder-32b

ttft27ms

p50 from the nearest warm region

Throughput318tok/s

Single stream, no batching

Context128K

Full window, no sliding truncation

Price$0.18

Per million tokens served

Fill-in-the-middle is supported, so this is the one to point an editor plugin at. Prefix and suffix are sent as separate fields rather than glued into one prompt.

Deploymentdep_45c48c

Parameters32B

Quantisationfp8

LicenceApache 2.0

Home regionsfo1

Extra replicas5

p50 latency29ms

p99 latency71ms

Requests / s2240

Figures are illustrative — this is a demonstration site.

Last 30 minutessfo1

cache hitqueue 0ms errors 0

Also in the catalogue

LL Llama 3.1 8B Instruct 19ms $0.06 MS Mistral Small 3 24B 22ms $0.14 PH Phi-4 14B 20ms $0.09 LL Llama 3.3 70B Instruct 38ms $0.34

Put the model next to the user.

One command, thirty-one regions, and an invoice that matches what you actually served.

Deploy a model Talk to an engineer