GPT-OSS 20B
A compact model contributing an additional perspective to every query.
Candidate responseYour workspace for application access and evaluated intelligence.
Create a key, connect your application, and let Inference coordinate the answers.
A single place to manage application access.
Keep your key in your server environment.
Start with a Python or TypeScript example.
Manage your signed-in identity and platform profile.
Three models run in parallel. Multiple available answers are evaluated; a single available answer is returned directly. Authenticate with your platform key.
Create and revoke up to 3 API keys per account.
A managed workflow for generation, review, and selection.
Start with a question. Add code, a language, or reference context.
One request calls three candidate models in parallel.
Three distinct perspectives address the same question and context.
Multiple available answers are evaluated; a single available answer is returned directly.
The strongest acceptable response is selected and delivered to your application.
Each model contributes a candidate response to the same request.
A compact model contributing an additional perspective to every query.
Candidate responseA larger model contributing a separate response to the same context.
Candidate responseAn additional model broadening the set of candidate responses.
Candidate responseCreate and revoke application keys with your signed-in account. Use a platform key to authenticate query requests.
Use the current HTTP contract in Python and TypeScript. Send a query and optional code with a platform API key.
Usage and history will show your actual query activity once those services are connected.