Models
Pick any model from the catalog. Run it in the Chat UI or over the API. Same experience, standard or confidential.
Run any model.Your way.
Chat UI or API. Standard or confidential. One studio to run inference, compare models, and deploy agents. No infrastructure to manage.
How it worksChoose a model.
Choose a model.
Run it your way.
01
Choose your LLM
From the catalog.
02
Chat UI or API
Same model, two interfaces.
03
Standard or confidential
Pick the right mode for the workload.
What you getModels, compare,
Models, compare,
APIs, and agents.
Compare
Send one prompt to multiple LLMs. Compare responses, latency, and cost before you ship.
APIs
OpenAI-compatible REST and streaming. Scoped keys, usage tracking, and governance out of the box.
Agents
Pre-built agents for coding and more. Deploy and customize without managing infrastructure.
Two modesStandard.
Standard.
Or confidential.
Standard
Fast, shared compute. The default for most workloads. Run any model in the Chat UI or over the API in seconds.
- Instant model access
- OpenAI-compatible APIs
- Broad model support
- Usage tracking and governance
Confidential
Attested, privacy-preserving GPU and CPU compute. Your prompts and outputs stay inside confidential environments. Built for regulated and sensitive workloads.
- Hardware attestation (GPU and CPU)
- Prompts never leave the environment
- Compliance-ready verification
Run inferenceat scale.
Browse models, compare outputs, and call the API. No infrastructure to manage.