# AI Model Price & Capability Monitor - Models.dev — workflow

## Choose this tool

AI engineering and FinOps teams maintaining a model-selection inventory. Provider/model identities connect price, context and capability changes.

Useful output columns: `providerId`, `modelId`, `inputUsdPerMillion`, `outputUsdPerMillion`, `contextTokens`, `outputTokens`.

## Practical workflow

Start with providers used by your application. Set a sufficient record window, review catalog prices against official provider quotes, and establish a baseline. Export changed price and capability fields to your model-selection review; this tool does not run benchmarks.

## Check the fit

Models.dev is a community-maintained catalog, not a provider quote or benchmark. Prices are listed USD per million tokens; missing values stay null and zero is preserved. Cached pricing, negotiated tiers and runtime availability need separate verification. No model inference calls are made.

## Real output sample

One selected row from a 2026-10-18 run. One saved row with selected scalar fields. Nested structures remain available in full Actor output.

Run: nyseux0cs6b7Ddm4K. Summary status: ok. Incomplete: false.

Historical bounded output from the recorded run. Current source values may differ.

```json
{
  "recordId": "anthropic/claude-fable-5",
  "title": "Claude Fable 5",
  "providerId": "anthropic",
  "modelId": "claude-fable-5",
  "inputUsdPerMillion": 10,
  "outputUsdPerMillion": 50,
  "contextTokens": 1000000,
  "outputTokens": 128000,
  "toolCall": true,
  "reasoning": true,
  "url": "https://models.dev/",
  "observedAt": "2026-10-18T15:33:27.289Z"
}
```