A tiny model playground

Qwen3.8
SAT Question Lab

Create an original SAT-style practice question, then inspect the answer and reasoning. Every result is an AI draft that should be checked by a person.

Qwen3.8-27B · Novita route

One-time setup: get your token

  1. Open Hugging Face token settings and choose Create new token.
  2. Choose Fine-grained. Under Inference Providers, enable Make calls to Inference Providers.
  3. Create and copy the token, then paste it below. Keep it private; this demo does not save it.
Needs “Make calls to Inference Providers” permission. It stays in this tab only and is not stored.
Billing: Hugging Face currently gives free accounts limited monthly inference credits. The listed price for this exact Novita route is unavailable, so this demo cannot promise a $0 call. Check your account billing before submitting. The request is pinned to Novita and will not fall back to another provider.

Your draft

Choose a skill and generate a draft. The response appears here; this Space does not save prompts or answers.

Model: Qwen/Qwen3.8-27B:novita. Your prompt and token are sent to Hugging Face's inference router to fulfill your request. The selected provider is Novita. This is not an official SAT product.

Token help: create a fine-grained Hugging Face token, grant Inference Providers permission only, and keep it private.