stack everyday.
English

Token prices(USD/1M tokens)

USD / 1M tokens
ModelInputOutputCached inputSource
gemini-3.8-flashGoogle $0.75$3.75$0.075Source Oct 10
gemini-3.6-flashGoogle $0.75$3.75$0.075Source Oct 10
gemini-3.5-flash-liteGoogle $0.30$2.50$0.03Source Oct 10
gemini-3.1-flash-liteGoogle $0.25$1.50$0.025Source Oct 10
gemini-3.1-pro-previewGoogle $2.00$12.00$0.20Source Oct 10
gemini-3.1-pro-preview-customtoolsGoogle $2.00$12.00$0.20Source Oct 10
gemini-3-flash-previewGoogle $0.50$3.00$0.05Source Oct 10
gemini-2.5-proGoogle $1.25$10.00$0.125Source Oct 10
gemini-2.5-flashGoogle $0.30$2.50$0.03Source Oct 10
gemini-2.5-flash-liteGoogle $0.10$0.40$0.01Source Oct 10
gemini-robotics-er-2-previewGoogle $1.00$5.00$0.10Source Oct 10

Standard API rates. Taxes, tools and subscriptions excluded.

Context

gemini-3.8-flash · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage. input: $0.75 through December 31, 2026. $1.50 starting January 1, 2027. output: $3.75 through December 31, 2026. $7.50 starting January 1, 2027. cachedInput: $0.075 through December 31, 2026. $0.15 starting January 1, 2027.

gemini-3.6-flash · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage. input: $0.75 through December 31, 2026. $1.50 starting January 1, 2027. output: $3.75 through December 31, 2026. $7.50 starting January 1, 2027. cachedInput: $0.075 through December 31, 2026. $0.15 starting January 1, 2027.

gemini-3.5-flash-lite · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage.

gemini-3.1-flash-lite · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage.

gemini-3.1-pro-preview · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage. Base prices apply to prompts <= 200,000 tokens; long prices apply to prompts > 200,000 tokens.

gemini-3.1-pro-preview-customtools · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage. Base prices apply to prompts <= 200,000 tokens; long prices apply to prompts > 200,000 tokens.

gemini-3-flash-preview · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage.

gemini-2.5-pro · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage. Base prices apply to prompts <= 200,000 tokens; long prices apply to prompts > 200,000 tokens.

gemini-2.5-flash · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage.

gemini-2.5-flash-lite · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage.

gemini-robotics-er-2-preview · Standard paid text rates, USD per 1M tokens. Output includes thinking tokens; excludes audio rates, tools and cache storage. input: $1.00 (text / image / video / audio) through December 31, 2026. $2.00 (text / image / video / audio) starting January 1, 2027. output: $5.00 through December 31, 2026. $10.00 starting January 1, 2027. cachedInput: $0.10 through December 31, 2026. $0.20 starting January 1, 2027.

Estimate token cost

Estimated cost—

Token-only estimate, not an actual task bill. Extra tool calls and retries are excluded.

Value for money →

Frequently asked questions

More
What unit are AI token prices quoted in?

Prices are US dollars per one million tokens. Input, output and cached input are separate rates. These are API prices, not ChatGPT, Claude or Grok subscription fees.

How is an API request cost calculated?

Multiply each token count by its matching rate, add the amounts, then divide by one million. Cached input replaces part of the input count; it is not counted twice. The calculator uses verified rates and shows unknown when a required rate is unavailable.

Does changing a model change the table order?

On the home comparison table, selecting a model updates only that provider row. Click an input or output column arrow to sort explicitly. A low input price does not necessarily mean a low total cost.

Can I assume every input token is billed at the cached price?

No. A listed cached rate does not guarantee a cache hit. Use cached billing only for input covered by the provider’s cache rules. The calculator requires cached input to be no greater than total input and does not assume a discount when its rate is unknown.

Can the token price change for a long context?

Some published pricing conditions distinguish context lengths or other request settings. Check the context notes and source for the selected model. The calculator uses a recorded long-context rate when available; it does not invent a missing rate.

Is the model with the cheapest input always cheapest overall?

No. Total cost depends on input, output and applicable cached tokens. A model with cheap input but expensive output may cost more for a long answer. Enter your expected token counts in the calculator and compare the rates separately.