Google released Gemini 4 Argon on October 13, 2026, and it’s now in both the Gemini app and the Gemini API. If you rewrite PLR, draft outlines, or churn out product descriptions in bulk, the details that matter are price, limits, and what you still have to do yourself.
Where you can use it
In the Gemini app, Argon is available to Google AI Pro and Ultra subscribers. According to Google, it gives more direct answers, hedges less, repeats itself less, and avoids clichés. Those are Google’s own claims, not independent test results. The app also has a new “tasks” mode that can keep working while you’re away from your computer.
On the API side, the model ID is “gemini-4-argon” and Google lists it as a stable model. It accepts text, images, video, audio, and PDFs, and it outputs text only. Its 1,048,576-token context window is shared between input and output. Default output tops out at 65,536 tokens, with a “long decode” option for more. It does not generate images or speech.
What it costs (as of October 2026)
Google’s pricing page lists introductory rates that run through November 30, 2026. Starting December 1, 2026, those rates double:
- Standard: $2 per million input tokens and $10 per million output tokens (thinking tokens count as output), rising to $4 and $20.
- Cached input: $0.10 per million, rising to $0.20.
- Batch and Flex: half the standard rates, so $1 input and $5 output, rising to $2 and $10.
- Free tier: none. Argon is paid-only on the API.
Google’s model page says Argon supports only implicit caching, not explicit caching, so you can’t manually set up a cache. If you plan a big PLR rewrite project, budget for the December prices, not the launch prices. Non-urgent bulk jobs are also a natural fit for the Batch rate.
Human editing is still on you
A stronger model doesn’t mean the output is ready to publish. Check facts, cut filler, and make rewritten PLR sound like your brand. This matters most for health, money, or legal topics, and for any claim in a product description.
Copyright and disclosure
Here’s some general information, not legal advice. In its January 2025 report, the U.S. Copyright Office said purely AI-generated material isn’t copyrightable, and prompts alone don’t make you the author. Your own creative edits, selection, and arrangement may be protectable, judged case by case. That matters if you plan to sell rights to AI-heavy content.
If you publish on Amazon KDP, its content guidelines require you to tell KDP about AI-generated text, images, or translations, even if you edited them heavily afterward. You don’t have to disclose AI-assisted work, meaning you wrote it and used AI to edit or brainstorm. Bulk PLR rewrites done by Argon will usually fall in the AI-generated category.
The bottom line
Argon is a capable paid option for high-volume text work, especially through the Batch rate. Run a small test batch, measure your real token use, and plan around the post-November pricing before you commit a whole catalog to it.
Sources
- Google Blog: Gemini 4 Argon availability in the Gemini app
- Google AI for Developers: Gemini 4 Argon model page
- Google AI for Developers: Gemini Developer API pricing
- Kindle Direct Publishing: Content Guidelines
- Skadden: Copyright Office Publishes Report on Copyrightability of AI Outputs
Featured image: Photo: StockSnap (CC0)