Updated
Gemini 4 Argon API pricing and the 1M-token answer
Gemini 4 Argon API pricing until 30 November 2026 is $2 per million input tokens and $10 per million output tokens. Cached input tokens are 95% off, so they cost $0.10 per million. From 1 December 2026 the rates rise to $4 and $20, and cached input to $0.20. One Argon answer can run to a million tokens. The output tokens of such an answer alone cost $10 until the end of November and $20 from December.
I price AI agent projects for clients, so from Google's price list I take two numbers: what an answer that uses the whole new limit costs, and who checks it afterwards. Argon has been generally available since 13 October. I have not tested it, so I am working from the price list.
Benchmarks, use cases and the timeline are in my overview of the announcement and what it means for companies. Here I stay with the money.
How much does Gemini 4 Argon cost in the API?
Google gave two prices. The introductory one applies until 30 November 2026: $2 per million input tokens, $10 per million output tokens and $0.10 for cached input. From 1 December 2026 the standard one applies: $4 and $20, with cached input at $0.20, so the 95% discount stays. In the 30 September announcement the footnote with the second price had no date. Today the Gemini API price list gives both dates. The Gemini 3.8 Flash promotion runs longer, until 31 December 2026.
| Model | Input | Cached input | Output | Max output (tokens) |
|---|---|---|---|---|
| Gemini 4 Argon, until 30 Nov 2026 | 2 | 0.10 | 10 | 65,536, up to 1M with long decode |
| Gemini 4 Argon, from 1 Dec 2026 | 4 | 0.20 | 20 | 65,536, up to 1M with long decode |
| Gemini 3.1 Pro Preview (prompts up to 200K tokens) | 2 | 0.20 plus storage | 12 | 65,536 |
| Gemini 3.8 Flash (promotion until 31 Dec 2026) | 0.75 | 0.075 plus storage | 3.75 | 65,536 |
| Claude Opus 5.5 | 4 | 0.20 (write 5) | 20 | 128K (300K in the Batch API, beta) |
| Claude Fable 5.1 | 10 | 0.25 (write 12.50) | 50 | 128K |
| GPT-6 Astra (prompts up to 272K tokens) | 10 | 1 (write 12.50) | 50 | 128K |
After the promotion, the closest match is Claude Opus 5.5, which costs exactly the same: $4 and $20, with cached input at $0.20. GPT-6 Astra and Claude Fable 5.1 cost $10 and $50. VentureBeat notes that at the introductory price Argon costs one fifth of GPT-6 Astra's price. After the promotion it is 40%.
In OpenAI's lineup, GPT-6.1 Sol has Argon's launch price: $2 and $10 in the API. After the promotion Argon will cost twice as much. I broke down the whole OpenAI price list after DevDay, from GPT-6 Luna to the ChatGPT Pro plans, in a separate piece.
What does one 1M-token answer cost?
Google raised the output token limit to one million, from the previous 64K. Current Gemini 3.x text models list a limit of 65,536 tokens in the docs. GPT-6 Astra, Claude Opus 5.5 and Claude Fable 5.1 stop at 128K tokens. Google explains the change with reasoning. A model with that much headroom can generate hundreds of thousands of tokens in one trajectory, think more deeply and solve a hard problem in one go. In the API the full million comes through long decode. The model then writes in hops of up to 64K tokens, and each hop picks up where the previous one stopped. Without long decode the limit in the Gemini API is 65,536 tokens. On Google Cloud a non-streaming direct HTTP request stops at 32,768 tokens.
The arithmetic is simple. One million output tokens times $10 per million is $10 at the introductory price, and after the promotion 1M times $20, which is $20. Then add the input tokens. With a 200K-token prompt you add 0.2 times $2, or $0.40, and $0.80 after the promotion. A full answer with that prompt therefore costs $10.40 today and $20.80 later.
One caveat. Thinking tokens count toward the limit, and Argon's price list bills them as output tokens. So the million tokens cover both the reasoning and the answer. You get a shorter text and pay for the whole thing.
On other models the same million output tokens take many calls. Claude Opus 5.5 charges $20 but needs at least eight calls, because 1,000,000 divided by 128,000 is 7.8. In the Batch API beta you get down to four, and Batch halves the price. GPT-6 Astra and Claude Fable 5.1 charge $50 each, also over at least eight calls. Gemini 3.1 Pro Preview charges $12, but with a 65,536-token limit it needs at least sixteen. Every call sends the prompt again, so you pay for it several times.
After the promotion, Argon and Opus 5.5 cost the same per token. Argon's gain is that the whole result comes out of one run and you do not have to stitch pieces together.
What fits in a million tokens of output?
Google does not convert tokens into pages or lines of code, so I work from an approximation. An average line of code has a few dozen characters, and a token is usually a few characters. I assume roughly 10 tokens per line as a working figure. It is my yardstick, Google gives none, and your language and tokenizer will give a different result.
On that yardstick a million tokens is on the order of 100K lines of code. A whole migrated module fits in one answer, tests included. Google writes that AI agents built on Argon are moving libraries of tens of thousands of lines to Rust, and a library like that fits. The Zircon kernel, which Google puts at over 800K lines, does not.
Text works the same way. A few characters per token gives a few million characters, which is over a thousand pages of 1,800 characters, minus whatever the thinking takes. A full report with appendices fits with room to spare.
One caveat on that number. Vals AI tested Argon before launch. Its model page lists a 262K-token output limit, and that is the setting it ran its tests with. The API docs now give two figures: 65,536 tokens by default and up to 1,048,576 with long decode, minus your input. The 1,048,576-token window is shared by input, thinking and output.
Designing or deploying agents for clients? In our collective of AI consultants you prepare for AI vendor partner certifications and learn with other consultants. When a suitable client brief arrives, we may invite you to a project.
Caching at 95% off: when does it pay?
With long prompts, caching moves the bill more than the answer limit does. Argon takes 95% off cached input. Current Gemini models take 90% off, for example $0.20 instead of $2 on Gemini 3.1 Pro Preview. Argon supports implicit caching only. When a request hits the cache, Google lowers the bill on its own. Explicit caching, where you create the cache yourself and pay to store it, is not available for Argon.
Take a company knowledge base: 200K tokens of policies and documentation placed at the start of every prompt. A hundred users each ask one question, so the model reads 100 times 0.2M, or 20M tokens.
Without caching you pay 20 times $2, or $40 at the introductory price. With caching the first prompt costs 0.2 times $2, or $0.40. The other 99 read 19.8M tokens from the cache at $0.10, or $1.98. In total you pay $2.38, which is 94% less. From December it comes to $80 without caching and $4.76 with it.
What does this bill include? Input tokens only, because you count the answers separately. I assume every one of the 99 following prompts hits the cache. With implicit caching that is optimistic. Google advises putting the shared content at the start of the prompt and sending such requests within a short time. I leave out a storage fee, because Argon's price list has none. Gemini 3.1 Pro Preview charges $4.50 per million tokens per hour to store an explicit cache, which for 200K tokens comes to $0.90 an hour.
| Model | Without caching | With caching |
|---|---|---|
| Gemini 4 Argon, until 30 Nov 2026 | 40.00 | 2.38 |
| Gemini 4 Argon, from 1 Dec 2026 | 80.00 | 4.76 |
| Gemini 3.1 Pro Preview | 40.00 | 4.36 plus storage |
| Gemini 3.8 Flash (promotion until 31 Dec 2026) | 15.00 | 1.64 plus storage |
| Claude Opus 5.5 (first write to a 5-minute cache) | 80.00 | 4.96 |
| Claude Fable 5.1 | 200.00 | 7.45 |
| GPT-6 Astra | 200.00 | 22.30 |
After the promotion, Argon and Opus 5.5 come out almost level here: $4.76 and $4.96. Caching pays when many prompts start with the same long block. That is what a knowledge base for a hundred users looks like, and so does an AI agent's fixed system prompt or the same tool definitions at every step. Keep the fixed part at the start and the user's question at the end. When every prompt carries a different document, the discount does little.
The introductory price trap: budget on $4 and $20
Building a product on Argon? Price it at $4 and $20. The promotion ends on 30 November 2026, so it covers at most the first weeks of your client contract. From 1 December you pay the full rate.
From 1 December input, output and cached tokens all cost exactly twice as much, so your model cost doubles too. The cache discount stays at 95%: $0.20 against a $4 price. The Batch API keeps half price: $1 and $5 until 30 November, then $2 and $10.
The price list also does not tell you how many tokens a model burns per task. Vals AI, which tested Argon before launch, measured that on its Vals Index at $4 and $20. The average cost per test came to $15.68 for Argon and $32.14 for Claude Opus 5.5, although both models have the same price per token. Vals measured on its own tasks, and yours may come out differently.
The Agentic Architect
Practical patterns, case studies, agentic projects taken apart and AI news. Zero spam, once a week.
You'll get one email to confirm. Unsubscribe with one click. Privacy policy
Who will read a million tokens?
After the promotion, a million output tokens cost $20. An engineer who has to read 100K lines of generated code will spend many days on them, and each of those days costs more than the model's whole answer. In my view that is where the real cost of a long answer sits. A million tokens nobody checked is a risk inside the client's system.
Google does not skip this either. The Rust migrations in the announcement go through automated and manual auditing, emulation testing and review before they reach production. Google does not say they are there yet.
The libgav1 example, Google's video decoder, gives a good pattern. AI agents built on Argon took an existing Rust port and replaced 32K lines of SIMD code. Google describes the result with two facts you can measure: the video output is identical, and the decoder runs 2.7x faster than the earlier Rust port. The baseline is that port, because according to Google the decoder only got closer to the C++ version. You check neither fact by reading the code line by line. That is what an acceptance condition for a long answer looks like.
Set one before you ask for a million tokens. It can be tests, matching results against the old system or a timing measurement. A second model that reads the result and approves it proves little. I wrote up separately a real bug that agreement between three AI agents would not have caught, and the check that rejects it.
Across a whole project budget the proportion looks similar. I showed it in a cost breakdown where checking and maintaining an AI agent after launch can outgrow the build cost within a year.
What did the launch price list settle?
The 30 September announcement left several gaps. The Gemini API price list and docs after the release close all of them.
- The introductory period runs until 30 November 2026. From 1 December the price is $4 and $20.
- Cached input keeps its 95% discount: $0.10 until 30 November, $0.20 from 1 December. Only implicit caching is available.
- The Batch API is 50% off: $1 and $5 until 30 November, then $2 and $10. Flex costs the same. Priority costs $3.60 and $18, and from 1 December $7.20 and $36.
- A long prompt does not raise the rate. Argon has one price on the list, without the 200K-token threshold Gemini 3.1 Pro Preview has.
- There is no fee for keeping data in the cache, because Argon does not support explicit caching.
- The price list bills thinking tokens as output, counted toward the limit.
- The model is called gemini-4-argon in the API.
- "1M" is a window of 1,048,576 tokens, shared by input, thinking and output.
One question stays open: how many tokens the model burns on your task. Only your own data will tell you that. I collected separately who got Argon first and what its availability looks like in Poland and the EU.
What to calculate before you move a project to Argon
- Recalculate the project budget at $4 and $20, because that is what you pay from 1 December. Count cached input at $0.20.
- Check in your logs how much of the prompt repeats between calls. That decides how much caching saves you.
- Decide how many output tokens you allow per task. The
max_output_tokenslimit includes thinking tokens too and defaults to 65,536 in the Gemini API. - Write down the acceptance condition for a long answer before you ask for one.
Then run your set of real tasks on Argon. Divide the whole bill by the number of tasks completed correctly and compare it with the model you use today.
Sources and fact-check date
- Google: Gemini 4 Argon announcement, with the price and the post-promotion footnote
- Google: the release post of 13 October 2026, with the footnote on the end of the promotion
- Gemini API: pricing, with the dates of Argon's prices
- Gemini API: what's new in Gemini 4 Argon, long decode and limits
- Gemini API: Gemini 4 Argon model page
- Gemini API: implicit and explicit caching
- Google Cloud: Gemini 4 Argon, output limits
- Gemini API: thinking tokens and max_output_tokens
- Gemini API: Gemini 3.1 Pro Preview model page
- Anthropic: Claude pricing
- Anthropic: Claude models overview and output limits
- OpenAI: API pricing
- OpenAI: GPT-6 Astra model page
- Vals AI: Gemini 4 Argon model page (262K output limit in its tests)
- Vals AI: Vals Index with cost per test
- VentureBeat: Google unveils Gemini 4 Argon
First published 1 October 2026. Argon's prices and limits rechecked on 23 October 2026.