↗ tokenulator an empathic tool Upload a session ↗

Compare model prices

A curated selection, not an exhaustive directory. OpenRouter selections follow weekly token usage; direct-provider families are curated, not a popularity ranking. Catalog metadata from Models.dev, with prices from OpenAI API pricing, Claude API pricing, Gemini API pricing, Kimi API pricing (batch rates), DeepSeek API pricing, Alibaba Qwen pricing, and OpenRouter's models API. New models are discovered daily. Estimates include published token and caching rates for the selected mode. Long-context surcharges, hourly cache storage, per-request fees, regional premiums, media/tool charges and taxes are excluded. Use these estimates only for requests within each model's base context pricing tier. Confirm pricing with the provider. OpenRouter figures are its advertised top-provider rates; actual routing and conditional pricing may differ. Gemini figures use the paid tier, not its free quota. Kimi figures are global USD API rates, not membership subscriptions or mainland China pricing. DeepSeek standard estimates use peak rates; select off-peak only for eligible hours in DeepSeek's current pricing terms. Qwen uses Singapore / International base-context prices for the text models featured on its official campaign page. Qwen cache rates remain unpriced; batch estimates apply only to explicitly eligible models.

Historical seed data adapted from AICostBudget, Model Price Watch, and AICostIndex under CC BY 4.0; Portkey under MIT. Normalized and filtered for text-token estimates. Dates on historical points are observation/verification dates unless explicitly marked as official releases. Secondary sources never replace current official rates.

© 2026 Empathic Inc.