AI Models, Pricing, and API Examples
Compare published AI models across text, image, video, audio, embeddings, and direct provider APIs.
Last verified: 2026-06-05
How to use the model catalog
The model catalog is the operational source of truth for published model IDs, providers, modalities, endpoint support, pricing summaries, context hints, workflow filters, and capability signals. Use it before sending traffic because display names may be prettified while API requests require exact model IDs.
- Use the top workflow tabs for broad jobs such as coding agents, creative work, automation, long-context analysis, and low-cost production.
- Use the left-side provider, modality, endpoint, and capability filters for exact narrowing before choosing a model.
- Open the model drawer to copy the exact API model ID and endpoint sample.
- Treat capability filters such as Cursor-ready, Cline-ready, Codex-ready, Open WebUI-ready, n8n-ready, Tool calling, Streaming, and Vision input as discovery signals, then verify support in the model drawer.
- Verify the first production request in Usage Logs before increasing volume.
From discovery to production
After choosing a model or tool, move through the same production path: copy the exact model ID, confirm endpoint support, set bounded parameters, send a short request, and inspect Usage Logs before scaling. This gives teams a repeatable way to compare cost, reliability, and request behavior before increasing traffic.
- Use /docs/api-reference for endpoint examples.
- Use /tools/ai-api-cost-calculator for rough cost planning.
- Email support@metisrouter.com with request IDs when a production issue needs help.
FAQ
Is this page the source of truth for API calls?
Use the model catalog, API docs, and Usage Logs together. Directory pages help discovery, while exact model IDs and endpoint support live in /models and /docs/models.
How should I compare candidates?
Run the same bounded task with each exact model ID, then compare latency, quality, failure rate, and billed usage.