INDEPENDENT GEMINI GUIDE
Estimate Gemini 4 Argon token costs
Understand Google's announced Argon introductory and later token rates, calculate a simple request estimate and separate pricing from API availability.
Sources checked . Availability and rankings can change.
The short answer
Google announced introductory rates of $2 per million input tokens and $10 per million output tokens, with cached input discounted by 95%. After the introductory period, the announced input/output rates are $4/$20. Checked October 8, 2026 (Beijing time).
A simple introductory-price estimate
For 100,000 uncached input tokens and 10,000 output tokens, the token-only estimate is $0.30: $0.20 input plus $0.10 output. At the announced post-introductory input/output rates, the same uncached example is $0.60.
Make the inputs match your workload
Use separate counts for uncached input, cached input and output. Count each input token once. Start with representative requests, then compare the estimate with actual usage when your account has access.
- The calculator shows introductory and later-price estimates side by side.
- Keep retries, ancillary services and any separate storage charges outside a token-only estimate.
- Budget a range when output length varies; a single optimistic example is not a monthly forecast.
A price is not an access guarantee
The reviewed announcement does not give an introductory-period end date. The official API model catalog reviewed for this guide did not list an Argon identifier. Confirm the current rate and project access before integrating.
Common questions
Does this site sell Gemini API access?
No. This is an independent guide and calculator; it does not provide Argon access or collect Google API keys.
Are estimates guaranteed invoices?
No. They illustrate token costs under the selected assumptions. Your provider's current terms and recorded usage determine actual charges.
Check the sources
Argon Budget is an independent resource, not a Google service. We do not provide model access or claim to have reproduced vendor benchmarks.