Selection
Model use cases
Use cases are selection workflows, not recommendations. Each one names the verified fields a reader should weight when screening models for a particular problem. WebmasterID Models does not declare a model 'best for' any use case.
Outcomes
Outcome-driven workflows
Outcome pages name the problem you are trying to solve and route you through the existing learn / apply / verify / test / package surfaces. Each ends with named Markdown artifacts — no recommendations, no rankings.
Outcome
AI model evaluation for developers
Evaluate AI models before integration without trusting a leaderboard
Hosted-provider mapping noteVerified comparison URLDecision evidence brief (Markdown)+2 moreOpen outcome →
Outcome
AI model selection for product teams
Turn a product use case into a defensible model selection
Use-case shortlist URLPricing-reference noteLifecycle risk note+1 moreOpen outcome →
Outcome
AI automation testing
Test an AI model inside an automation before it runs unattended
Automation risk checklistPrompt test matrixSafe model-use checklist+2 moreOpen outcome →
Outcome
AI model governance review
Build a defensible governance review with sourced evidence
Source freshness checklistLifecycle review noteData gap list+2 moreOpen outcome →
Outcome
LLM prompt evaluation
Evaluate model behaviour with safe, structured prompt sets
Per-set Markdown export (via /api/lab/prompts/<slug>)Prompt test matrix (Markdown)Per-prompt observation record+1 moreOpen outcome →
Outcome
Structured output testing
Validate JSON mode, structured output, and tool calling against your real schema
Schema bytes captured alongside resultsPer-prompt validator outputField-level failure mode list+1 moreOpen outcome →
Detail pages
Use cases with detail pages
Each detail page walks the verified fields, the most common misreads, and the source-backed shortlist.
Long-context analysis
Working with prompts and corpora that exceed the small-context tier (≥200k tokens). The relevant verified fields are context window, max output tokens, and the pricing reference rows that apply at the prompt-size tier the vendor publishes.
Verified fields used: context window · max output tokens · pricing references (incl. >200k tier where the vendor publishes one) · …
Multimodal input
Workloads that pass images, audio, video, or PDFs in addition to text. The verified modality channel list is the only safe filter — unsupported modalities remain explicitly unverified.
Verified fields used: modality channel list (input + output) · context window · lifecycle status · …
Hosted inference
When the model creator does not run a paid first-party API and inference is delivered by a third-party hosting platform (Groq, Together AI). Hosted availability is a stable identity claim; hosted pricing is a volatile reference value.
Verified fields used: hosted availability record (host × model) · hosted model ID · billing provider slug · …
Governance review
Internal AI model inventory + source-backed due diligence. The relevant signals are verification status, lifecycle, source citations, freshness states, and the reverification queue — all derived from primary-source documentation, with explicit data gaps where the vendor does not publish.
Verified fields used: verification status · lifecycle status · source citations with retrievedAt · …
Reserved
Use cases — deep dive coming
The slugs below are reserved so the catalogue can grow without breaking links. They describe selection workflows the data layer already supports, but no dedicated detail page ships this sprint.
Structured output
Workloads that depend on JSON, tool-calls, schema validation, or other structured response surfaces. The catalogue does not yet model structured-output guarantees as a verified field; this use case currently surfaces models with verified tool-use signals where the vendor docs publish them.
Verified fields used: features (tool use, where verified) · modality output channels · source citations
Workflow currently available via /select?useCase=structured-output.
Cost review
Walking pricing references before a procurement decision. The catalogue exposes first-party rates as source-backed references (not live quotes) alongside hosted-provider references; the freshness chip tells you how recently each was confirmed.
Verified fields used: first-party pricing references · hosted pricing references · pricing freshness state · …
Workflow currently available via /select?useCase=cost-review.
Status-aware selection
Picking models whose providers expose either a vendor-reported status page or an independent host probe. The catalogue records observations, not uptime; readers can use the status surface to verify *that monitoring exists*, not to assert SLA compliance.
Verified fields used: provider statusPageUrl · wired observer presence · status observations (vendor-reported / independent probe)
Workflow currently available via /select?useCase=status-aware-selection.
Comparison research
Reading two models side-by-side after the use case has narrowed which fields matter. Comparison pages render verified fields per side — they do not declare a winner and do not synthesise derived metrics.
Verified fields used: every verified field on each comparison's two model records · comparison cluster membership
Workflow currently available via /select?useCase=comparison-research.