Guided demo
Walk a hosted-inference workflow where the model creator does not run a paid first-party API. Pick the use case, narrow the shortlist to models with verified hosted availability, then compare hosted pricing references side by side. The brief separates the hosting platform (billing provider) from the model creator at every step.
Workflow preview
Server-rendered from current catalogue data. Each tile links to the live surface the demo walks through.
Step 1 · use case
Hosted inferenceNames the verified fields this workflow weights.
Step 2 · shortlist
2 candidatesDeterministic order — no scoring.
Step 3 · comparison
2 models side by sideVerified fields only — no derived metrics.
Step 4 · brief
5 evidence fieldsMarkdown or JSON. Evidence, not recommendation.
Step 5 · sources
Citation registryEvery value linked back to a primary source.
Hosted pricing is set by the hosting platform — not by the model creator. The demo does NOT rank hosting platforms by price. Always re-verify the platform's own pricing page before commitment.
Steps
Follow the route plan below. Each step links to the live surface and explains what to inspect there.
Step 1
Read the use-case guide
Name which verified fields matter for this workflow before picking candidates.
/use-cases/hosted-inference
Step 2
Open the selection workspace
Narrow a source-backed shortlist with the use-case filter pre-applied.
/select?useCase=hosted-inference&hostedAvailability=true
Step 3
Compare verified fields
Render the selected models side by side from the verified data layer.
/compare/build?models=llama-4-scout%2Cdeepseek-v4-pro&useCase=hosted-inference
Step 4
Generate an evidence brief
Export a paste-ready evidence pack (Markdown or JSON).
/briefs/build?models=llama-4-scout%2Cdeepseek-v4-pro&useCase=hosted-inference
Step 5
Inspect sources + freshness
Open every primary-source citation referenced by the selected models.
/sources
Evidence preview
Rendered for the 2 models included in this demo. Every value comes straight from the typed local data layer.
Hosted inference — verified fields
| Model | Lifecycle | Context | Modalities | Hosted | Sources | Freshness |
|---|---|---|---|---|---|---|
| Llama 4 Scout | active | 10,000,000 | text-in, image-in, text-out | groq | 1 | Fresh |
| DeepSeek V4 Pro | active | 1,000,000 | — | together-ai | 3 | Fresh |
Preview generated from current catalogue data. Every value is rendered straight from the verified data layer; unknown values stay null.
Decision brief preview
Server-rendered from the same helper that powers /briefs/build and /api/briefs/decision. Markdown + JSON export links are live.
Hosted inference — brief preview
Generated at: 2026-05-24 · Use case: Hosted inference
Models
2
Evidence rows
14
Data gaps
7
Sources
6
Preview generated from current catalogue data. Briefs are evidence, not recommendations.
What this demo does not decide