← All tools

API costs

AI model cost calculator

Compare text API costs across six providers, including token usage, caching and request fees.

Reviewed · Free to use · Calculations run in your browser

Your usage

Token usage

Caching, rates & additional costs

Cache reads and writes are portions of total input, not extra tokens. Leave both at zero for uncached usage.

Edit rates for Batch, different cache durations, regional processing or your own contract. Edited rates become a custom estimate. Other charges can include search tools, cache storage, images and audio.

How this calculator works

Uncached input, cache reads, cache writes and output are priced separately, divided by one million and multiplied by the request count. The request fee and any additional costs are then added. Use the same period for requests and additional costs.

Presets are a reviewed selection, not a complete model catalog or a live billing feed. Standard text inference rates are shown unless the model label says otherwise. Use your provider's measured token counts: the same text can tokenize differently across models.

OpenAI presets include the published short- and long-context bands, with 30-minute cache-write pricing; regional uplifts and Fast mode are excluded.

Anthropic cache writes use the five-minute rate. DeepSeek peak hours are weekdays 01:00–04:00 and 06:00–10:00 UTC; other hours use off-peak rates. Perplexity presets include standard-search request fees; Pro Search and Deep Research have additional billing rules.

Sources & review date

Checked September 11, 2026. Providers can change rates, model availability and specifications. Open the source for the selected preset before budgeting or buying hardware.