Trainium Demo

Try models

Model

Speculative Decoding

Trainium preparation is in progress for this model. Please check back soon.

Ending demo

Stopping both inference engines…

Model selection will return when both hosts confirm shutdown.

Preparing

Preparing

trn2-2 · Qwen3.8-27B

TTFT — · — tok/s · — tok/$

trn2-4 · Base / DFlash

TTFT — · — tok/s · — tok/$

0 / 200 words

Select an example, then choose Test to fill the input. Send starts generation.

vLLM

llm-neuron