Input tokens rise from $2 to $4 per million and output tokens from $10 to $20 – UK developers using the Gemini Developer API will feel it too.
Key points
- Google is raising the standard price for one million input tokens from $2 to $4.
- One million output tokens will cost $20 instead of $10; thinking tokens are included in that count.
- The change applies to the paid Gemini Developer API for Gemini 4 Argon. The model is not included in the free API tier.
From 1 December 2026, Google is doubling the standard prices of its Gemini 4 Argon AI model in the Gemini Developer API. One million input tokens will then cost $4 instead of $2, and one million output tokens $20 instead of $10. The new price list applies to applications that call the model through the Gemini Developer API – it is not an announcement of higher monthly fees for the Gemini app.
Google bills Gemini Developer API usage worldwide in US dollars, regardless of the developer’s country. That means the increase hits teams in the UK exactly as it hits teams in the US or Germany: there is no separate British price list, and currency conversion happens only when a card or invoice is settled, at whatever rate the payment provider applies on that day.
The new API prices
Google’s Gemini Developer API pricing page currently lists $2 per one million input tokens and $10 per one million output tokens for Gemini 4 Argon. According to the table, these rates apply up to and including 30 November 2026. From 1 December, both rise to $4 and $20 respectively – a 100 percent increase on each token type.
| Gemini 4 Argon, standard tier | Until 30 November | From 1 December |
|---|---|---|
| Input, per 1 million tokens | $2 | $4 |
| Output, per 1 million tokens | $10 | $20 |
| Context caching, per 1 million tokens | $0.10 | $0.20 |
The listed context-caching price doubles as well. Google marks only implicit caching as available for this model. The table covers the standard tier; batch processing has its own price tier. Google’s pricing page does not list Gemini 4 Argon as available free of charge.
Sample calculation
- Input: 1 million tokens × $2 (until 30 Nov) / $4 (from 1 Dec)
- Output: 100,000 tokens × $10 (until 30 Nov) / $20 (from 1 Dec) per million, i.e. $1 / $2
Total for this simplified call: $3 until 30 November, $6 from 1 December. This is a model calculation without Google Search grounding, caching, discounts, tax or currency conversion – not a forecast for any specific invoice.
Google explicitly counts the model’s thinking tokens as part of the output total. If an application uses function calling or Google Search grounding, additional tokens or separate search queries can be billed on top; the documentation notes that a single request can trigger one or more search calls. Developers in the UK and elsewhere should check the billing breakdown of their actual API calls rather than just the visible answer length.
What developers should watch
According to Google’s model overview, Gemini 4 Argon is built for demanding tasks such as complex software development, multi-step agents and large context windows. That can enable powerful applications, but it also makes the cost per request strongly dependent on input length, output length and any additional tool use.
Checklist before 1 December
- Pull actual token usage for your production workloads from the Google Cloud or AI Studio billing dashboard.
- Recalculate typical workflows with the new $4 / $20 rates, including thinking tokens in the output count.
- Check whether a shorter context, batch processing or a different model tier fits the required quality and latency.
- Confirm with your finance team whether the increase changes a UK-based budget set in GBP, since billing stays in USD.
Google introduced Gemini 4 Argon in October, for both the Gemini app and developers. App availability for Google AI Pro and Ultra subscribers is described separately from the API prices, so the new API table does not by itself indicate any change to a monthly app subscription. For API projects, the price list Google publishes at billing time is what applies.
Frequently asked questions
From when do the higher Gemini 4 Argon API prices apply?
Google names 1 December 2026 as the start date. According to its pricing table, the current standard prices apply up to and including 30 November.
Which prices are changing?
In the standard tier, input rises from $2 to $4 per one million tokens. Output rises from $10 to $20 per one million tokens. Context caching rises from $0.10 to $0.20 per one million tokens.
Is Gemini 4 Argon free to use?
In the Gemini Developer API, Google does not list the model as available in the free tier. The Gemini app has its own separate access and subscription rules.
Does a Google AI subscription also get more expensive?
The published change sits in the developer API’s price list. It does not document any increase to a monthly Google AI app subscription.
Does your team already use Gemini 4 Argon in a live application? Which cost line matters most for you under the new price list? Share your view in the comments.
Sources and status
- Google AI for Developers: Gemini Developer API – pricing, including Gemini 4 Argon, accessed 2 November 2026.
- Google AI for Developers: Gemini 4 Argon model overview, model update October 2026.
- Google: Gemini 4 Argon comes to the Gemini app, 13 October 2026.
Research and data as of 2 November 2026. Prices in US dollars, per one million tokens, excluding tax and any further fees.
Transparency note: AI-assisted tools were used in researching, structuring and writing this article. The sources used are listed in the article. Publication takes place only after manual editorial approval. Translation of the German original on Net Trends. The featured image is an infographic created specifically by the Net Trends editorial team.
Share this article
