Inference

Mixtral-8x7B Inference with SageMaker

Enter a prompt below to get a response from the model.

Runtime request path

Next.js
Route Handler
API Gateway
HTTPS
Lambda
Invoke
SageMaker
Endpoint
Mixtral
8x7B + adapter
Response
Normalized

Enter your prompt

Model response

No output yet

Generation parameters

Fixed by the deployed Lambda

The handler sets these values and applies them to every request.Also set server-side: do_sample true, top_k 50, return_full_text false, stop sequence </s>.

Full engineering case study — problem, approach, architecture, decisions