Gemini 4 Argon API pricing: $2/$10, then $4/$20

According to APITariff, Gemini 4 Argon API pricing is $2.00 per 1M input tokens and $10.00 per 1M output tokens on Google’s pricing page, read 2026-10-14, and that rate ends on 2026-11-30. From 2026-12-01 the page lists $4.00 and $20.00, twice the launch rate. Our fetch of OpenRouter at 07:10 UTC (Coordinated Universal Time) on 2026-10-14 first read the model at $2 and $10 with a 1,048,576-token context window. The reseller kie.ai lists it at $0.80 and $4.00. At 1,000,000 calls of 1,000 input and 300 output tokens that is $5,000 on Google now, $10,000 from December 1 and $2,000 on kie.ai.

Last updated 2026-10-14 · 6 models compared on input, output, context and bill · source: Google pricing page, APITariff fetch of OpenRouter and kie.ai, 2026-10-14.

Gemini 4 Argon vs 5 nearby models, field by field

We compared Gemini 4 Argon with the 5 text models nearest to it in price. At today’s rate it matches Claude Sonnet 5.5 and GPT-6.1 Sol to the cent and costs less per output token than Gemini 3.1 Pro. At the December rate it matches Claude Opus 5.5.

Model	Input / output $ per 1M	Context	Bill at 1M calls
Gemini 3.8 Flash	0.75 / 3.75	1,048,576	$1,875
Gemini 4 Argon, to 2026-11-30	2.00 / 10.00	1,048,576	$5,000
Claude Sonnet 5.5	2.00 / 10.00	1,000,000	$5,000
GPT-6.1 Sol	2.00 / 10.00	1,050,000	$5,000
Gemini 3.1 Pro (preview)	2.00 / 12.00	1,048,576	$5,600
Gemini 4 Argon, from 2026-12-01	4.00 / 20.00	1,048,576	$10,000
Claude Opus 5.5	4.00 / 20.00	1,000,000	$10,000

Sources: Google’s pricing page, read 2026-10-14, and APITariff’s OpenRouter fetch of 2026-10-14 07:10 UTC, no row quarantined. APITariff compares price only, with no quality score. The Claude Opus 5.5 post has that model’s rates in full, and the Gemini 4 Argon model page carries the daily reading.

What does Gemini 4 Argon cost at 10k, 100k and 1M calls?

At 1,000 input and 300 output tokens per call, Gemini 4 Argon costs $50 for 10,000 calls, $500 for 100,000 and $5,000 for 1,000,000 until 2026-11-30, then $100, $1,000 and $10,000. Batch halves the standard rate on Google’s page: $2,500 for 1,000,000 calls to November 30 and $5,000 after. Cached input is $0.10 per 1M tokens to November 30 and $0.20 after, and Google lists implicit caching only. Put your own token counts into the token calculator.

Is Gemini 4 Argon cheaper through kie.ai?

Yes. kie.ai, a reseller, lists Gemini 4 Argon at $0.80 input and $4.00 output per 1M tokens in its pricing API, fetched 2026-10-14 at 16:02 UTC. That is 60% under Google’s $2 and $10 and 80% under the $4 and $20 that apply from December 1, if kie.ai holds its rate; it has not published a December price. At 1,000,000 calls the bill is $2,000 against Google’s $5,000 now.

Model	kie.ai in / out $ per 1M	Google or vendor in / out $ per 1M	Bill at 1M calls via kie.ai
Gemini 4 Argon	0.80 / 4.00	2.00 / 10.00	$2,000
Claude Sonnet 5.5	0.80 / 4.00	2.00 / 10.00	$2,000
GPT-6.1 Sol	0.60 / 3.00	2.00 / 10.00	$1,500
Gemini 3.1 Pro	0.50 / 3.50	2.00 / 12.00	$1,550
Gemini 3.8 Flash	0.225 / 1.125	0.75 / 3.75	$562.50

Sources: kie.ai pricing API, fetched 2026-10-14 16:02 UTC, and OpenRouter, fetched 2026-10-14 07:10 UTC. kie.ai bills in credits; the dollar figures are its own conversion. kie.ai is not the cheapest route in every pairing: on kie.ai, GPT-6.1 Sol bills $500 less than Argon and Gemini 3.8 Flash bills $1,437.50 less. kie.ai is not Google, and we have not sent a Gemini request through it or read its data-retention terms. The kie.ai comparison of 31 models shows where its rates sit against list prices.

Affiliate link: if you sign up through it we may earn a commission. The link changed none of the figures above. kie.ai pricing lists the full rates; our kie.ai price page goes further.

Is Gemini 4 Argon free or open source?

Gemini 4 Argon is not free. Google’s pricing page, read 2026-10-14, lists the free tier as “Not available” for the model and says it is not available on the Google AI Pro plan. Paid-tier requests are not used to improve Google’s products, per the same page. The page says nothing on open weights or a waitlist.

How we compared

We took Google’s own pricing page as the list price, paired Gemini 4 Argon with the nearest models by price from our fetch, and priced each at 1,000 input and 300 output tokens per call. We have no launch date: we read the model first in our fetch of 2026-10-14, and the page we read carries no date. OpenRouter’s row carries one rate, $2 and $10, so it will show no December change until the provider lists one. Collection details are on the methodology page, and the cheapest LLM API ranking puts these rows beside every text model we fetch.

Cite as: According to APITariff (apitariff.com/news/gemini-4-argon-api-pricing/, read 2026-10-14).

Drafted with AI from APITariff’s own price data, checked and edited by a person. A script draws the chart from the figures in the tables above; no image is generated.

Change log

• 2026-10-14First edition. Google rates read on its pricing page 2026-10-14; kie.ai row fetched 2026-10-14 16:02 UTC.