AI API Cost Calculator - Ai-Forum.org

Choose a model Reference model

All prices are editable. Changing models restores that model’s reference prices.

Your workload

Total input tokens

Per request. Include instructions, history, retrieved context, and cache hits.

Output tokens

Per request. Include billable reasoning or thinking tokens.

API requests per day

Count every model call, including additional workflow steps and retries.

Active days per month

Use 30 for every day, or your expected active days (1–31).

Token prices

Input price (USD / 1M)

Output price (USD / 1M)

Optional: cached input

Cache hits

Cached input tokens

Per request. A subset of total input; use 0 if you do not expect cache hits.

Cache hit price (USD / 1M)

Cache read / hit rate only. Cache creation and storage fees are excluded.

How the estimate works

Cost per request = ((total input − cached input) × input price + cached input × cache hit price + output × output price) ÷ 1,000,000. Monthly cost = cost per request × daily requests × active days.

Assumes the same average token counts and cache hits for each request. Use actual billable token counts, not word counts. This estimates token charges only: taxes, subscriptions, tool calls, cache creation/storage, audio/image/video billing, routing surcharges, tier changes, discounts, and free quotas are excluded. Presets are dated reference prices, not a live price feed. Verify your provider’s pricing tier and model availability before budgeting.

Calculated in your browser. No API key, sign-in, or saved input required.