Llama 3.3 70B Instruct

The default for anything that has to sound like a person and still be checkable.

$argosy up --model llama-3-3-70b-instruct

ttft38ms

p50 from the nearest warm region

Throughput212tok/s

Single stream, no batching

Context128K

Full window, no sliding truncation

Price$0.34

Per million tokens served

The default choice for anything that needs to sound like a person and still be checkable. Serves out of every region; we keep six warm replicas on the US east and west coasts because that is where the traffic is.

Deploymentdep_167909

Parameters70B

Quantisationfp8

LicenceLlama 3.3 Community

Home regioniad1

Extra replicas6

p50 latency41ms

p99 latency96ms

Requests / s1420

Figures are illustrative — this is a demonstration site.

Last 30 minutesiad1

cache hitqueue 0ms errors 0

Also in the catalogue

LL Llama 3.1 8B Instruct 19ms $0.06 QW Qwen2.5 72B Instruct 41ms $0.36 MS Mistral Small 3 24B 22ms $0.14 DR DeepSeek R1 Distill 32B 34ms $0.20

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