Get started
What is available
Prepared model artifacts, available browser workflows and released customer services establish different facts. This guide explains how to interpret the two Qwen candidate profiles, fictional deployment states and sample results without turning a prepared model profile into a public capacity claim.
SAVRN Cloud · Coming soon · Public preview · Browser-local simulation · No live compute or payments
Read status in its proper context
The SAVRN Cloud interface is available as a public browser preview. Connected inference, compute, evaluation and training services remain Coming soon. A working screen, a prepared model artifact and an available customer service are different states; each needs different evidence.
| Status or record | What it establishes here |
|---|---|
| Prepared Qwen candidate | A specific configuration is documented for further qualification. |
| Ready demonstration deployment | A fictional browser record can support its sample workflow. |
| Completed sample | A deterministic fixture produced a local result and receipt. |
| Connected model service | Not available through this public preview. |
Understand the two candidate profiles
Qwen3.5-9B Q6_K and Qwen3.8-27B Q5_K_M are prepared configurations for SAVRN-operated inference. Their utility and coding roles describe intended evaluation tasks. No SAVRN quality ranking, throughput, customer price or available capacity is published for them. The proposed 8,192-token context and 1,024-token output bounds still need qualification.
An independently allocated serving lane must be approved before either candidate can become a connected service. A prepared profile does not establish available customer capacity.
Interpret a playground result
The guided starters prepare requests for evidence review, extraction and research code. The receipt distinguishes an intended candidate from the fictional fixture that actually supplied the browser sample. No Qwen execution, external source fetch or code execution occurs. Sample usage and Work Credits do not represent a bill or a production token measurement.
Research 8B begins with a ready fictional deployment. Other runnable demonstration profiles may need a demo deployment before a sample can run. A disabled action should explain its unmet preview requirement; it is not evidence of hidden capacity.
Before a service can launch
A release requires exact runtime identity, task evaluations, supported-feature tests, measured performance, independently allocated capacity and enforceable access. License eligibility, pricing and accounting require their own approvals. Inspect Roadmap and Platform map for the intended sequence. A successful local walkthrough proves only that its demonstrated workflow completed.
Build toward the work that matters.
Tell SAVRN what your institution needs to run, who reviews the results and where its data must stay. That workload defines the next service to qualify.
Discuss an AI project