LLMRates

Features

Everything in one reconciled API: model catalogue, history, recommendations, natural-language query endpoints, stream updates, and webhook automation.

Core Pricing APIPRICING
GET/v1/models
Full model catalogue with current prices and metadata
GET/v1/models/:id
Model detail, full price history, and source attribution
GET/v1/providers
Provider catalogue
GET/v1/compare
Side-by-side model comparison
GET/v1/changes
Recent price deltas with source and confidence
History & ContextHISTORY
GET/v1/models/:id/history
Full price timeline — every change, every source
GET/v1/context
~2k token pricing snapshot for agent system prompts
GET/v1/recommend
Ranked model recommendations by task, context, and price
Agent & AutomationAGENT
POST/v1/ask
Natural language + structured pricing response
GET/v1/stream/changes
SSE stream with Last-Event-ID reconnection semantics
POST/v1/webhooks
Signed webhook subscriptions for price change events
DELETE/v1/webhooks/:id
Webhook removal

Frequently Asked

+How are prices verified?

Data is scraped from multiple sources (OpenRouter, LiteLLM, provider docs) and reconciled. Divergent values >5% are flagged for manual operator review to ensure zero guesswork.

+How fast do changes appear?

Price updates are ingested continuously and usually surface through API endpoints and the SSE stream within 60 seconds of confirmation.

+Can I use this in production agents?

Yes. Our /v1/context and /v1/ask endpoints are optimized for prompt-injection and low-latency agent tool-use, returning lean JSON snapshots.

+What kind of questions can I ask the /v1/ask endpoint?

You can ask natural language questions like 'Cheapest model for summarization?', 'Models under $1/M input?', or 'Which providers dropped prices recently?'.

+How do GPT-4o and Claude 3.5 pricing compare?

GPT-4o is currently more aggressive on input pricing ($2.50 vs $3.00 per 1M tokens), while output costs are highly competitive across both. Use our comparison tool for live deltas.

+Are webhook payloads signed?

Yes. Webhook events are signed with HMAC-SHA256 using your secret key so recipients can verify they originated from LLMRates.

+What do 'input' and 'output' prices mean?

Input price is the cost per 1M tokens sent in your prompt. Output price is the cost per 1M tokens generated by the model in response.