Google released its flagship model Gemini 4 Argon on October 13, and the API pricing comes with an expiration date: $2 per million input tokens and $10 per million output tokens until November 30. From December 1, every tier doubles. If you're planning to build on it, the next six weeks are the cheapest they'll ever be. Here's the full price table, where it's available, and what Google's claims are actually worth.
What launched
Gemini 4 Argon (gemini-4-argon) is Google's new flagship, positioned for "deep work" — complex, multi-step tasks across coding, finance, law, cybersecurity, and science. It's generally available as of October 13 through the paid Gemini API. Two things to note upfront: it is not on the API's free tier, and this launch follows a messy stretch — Google scrapped the planned Gemini 3.5 Pro in September and overhauled DeepMind's leadership after Demis Hassabis stepped aside. This model has a lot to prove.
The full price table
Google published per-tier pricing with an introductory period ending November 30, 2026:
| Tier | Input / 1M | Output / 1M | Cached input / 1M | From Dec 1 (in / out / cached) |
|---|---|---|---|---|
| Standard | $2.00 | $10.00 | $0.10 | $4.00 / $20.00 / $0.20 |
| Batch and Flex | $1.00 | $5.00 | $0.05 | $2.00 / $10.00 / $0.10 |
| Priority | $3.60 | $18.00 | $0.18 | $7.20 / $36.00 / $0.36 |
A few things worth knowing: output prices include thinking tokens (the model's internal reasoning, which you pay for). Cached input — text the model has already processed — gets a 95% discount, which is unusually generous and matters a lot for agentic workflows that re-read context. Batch and Flex trade speed for half price. Priority costs more for faster, more reliable responses.
For context, Gemini 3.8 Flash sits at $0.75 input / $3.75 output on the same page — Argon is roughly 2.7x the price of Flash at intro rates, 5.3x after December.
Where you can actually use it
-
Paid Gemini API and Google AI Studio (
gemini-4-argon) - Antigravity, Google's agentic coding environment (requires Ultra subscription for individuals)
- Gemini Enterprise subscriptions
- Gemini app and AI Mode in Search (Pro and Ultra subscribers)
- Android Studio and Firebase AI Logic (Android developers)
The API has a unified limit of one million input-output tokens, with a continuation protocol for longer tasks.
The claims worth scrutinizing
Google says teams using Argon in Antigravity increased agentic code submissions by more than 35% while reducing rollbacks by nearly a third. Early enterprise testers include Deloitte, LexisNexis, Ambience Healthcare, and Isomorphic Labs. These are Google's numbers, from Google's own tool, with no independent benchmark published yet. Treat them as marketing until third parties verify.
Google is also opening its Fairwind Program to hundreds more cybersecurity partners — vetted defenders get a version of Argon without its cyber guardrails — and offering 10,000 scientists six months of free access to Antigravity's top tier. The scientist giveaway is a smart distribution play: get the model into labs before the competition does.
Should you build on it now?
The honest math: if your workload fits, the intro pricing window (until November 30) is real value — Standard tier at $2/$10 undercuts most frontier models. But architect for the December price ($4/$20), because that's the price you'll actually live with. The 95% cached-input discount is the sleeper feature: for agents that re-read long contexts, effective costs could land well below sticker price.
The risk is Google's recent track record — a scrapped 3.5 Pro and leadership turmoil don't inspire confidence in roadmap stability. Prototype now while it's cheap; commit after independent benchmarks land.
FAQ
Is Gemini 4 Argon on the free tier? No. Paid API only.
When does the intro pricing end? November 30, 2026. Standard rates apply December 1.
What's the cheapest way to try it? Batch tier at $1/$5 per million tokens during the intro period — if you can tolerate slower responses.
How does it compare to Claude and GPT pricing? At intro rates it's competitive with frontier pricing; after December it lands in the same band. The cached-input discount is more generous than most competitors, which matters for agentic use.
Sources: Reuters (Oct 13, 2026), Google's Gemini API pricing page and release notes, NeoTeo. Pricing figures are Google's published rates; performance claims are Google's and unverified independently.
Top comments (0)