AI Gateway · Component · task-model-routing-evaluation
Task-aware model routing and structured evaluation
What it does
Select configured GPT-family task tiers with ordered fallback and validate structured outputs against deterministic fixtures. Logical tiers map to configured models; this is not a multi-vendor billing gateway.
Core pattern
Task-tier fallback + fixture checks for structured results
- Input
- Prompt, task contract and expected evaluation cases
- Output
- Validated model payload or failure and evaluation results
Key parameters
| Name | Type | Default | Values |
|---|---|---|---|
| task_type | string | — | — |
| prompt | string | — | — |
| provider_order | array | [] | — |
| model_mapping | — | — | — |
| output_schema | object | {} | — |
Maturity and limits
Active router supports GPT model tiers; legacy provider names normalize to those tiers; harness/fixtures included. Source paths were inspected statically; tests and live integrations were not executed. This rating describes implemented code, not production validation.
Credential-bearing configuration is excluded from this inventory. No root reuse license found; external content/platform terms and internal hotel/business data access constrain reuse.
1 sentence citing source paths or vendor names omitted for publication.
Used in pipelines
Industry tags
- Travel & hotelsSource-supported industry
For work involving destinations, room rates and bookings, use task-tier fallback followed by fixture checks for structured results.
- EducationTransferable use
For sample acceptance checks before releasing a course question assistant, use task-tier fallback followed by fixture checks for structured results.
- Business & general servicesTransferable use
Support “Compare shared question samples” with task-tier fallback followed by fixture checks for structured results.
Skills pack
Contract scaffold and deterministic mock
Download · 12KBSHA-256 b52a491e4a55474c5d8a1a6c1096c27f8b035d8927492363c5cc0315d21a3f23
Related components
-
Multi-protocol model routing, fallback and compatibility
Request protocol normalization + compatible routing and fallback
Working 1 pipeline -
Model comparison and shadow evaluation
Shared-sample model comparison + quality and cost records
Working 1 pipeline -
Provider batch inference
Batch request submission + job state and result archives
Working 1 pipeline -
AI endpoint management and testing
Service configuration and connectivity tests + response inspection
Prototype 2 pipelines