What Gemini 4 Argon costs per million tokens, next to Claude and Grok

Google's October 13 post gives Gemini 4 Argon an introductory API price of $2 per million input tokens and $10 per million output tokens. Cached input is 95% off. The intro runs through November 30, 2026. After that, Google says the price is $4 per million input tokens and $20 per million output tokens. The context limit it describes is 1 million unified input and output tokens.

Who gets the model without using that API price: Google AI Ultra in Antigravity, and Google AI Ultra or Pro in the Gemini app and AI Mode in Search. US subscription prices checked the next day on gemini.google/subscriptions are Plus $4.99, Pro $19.99, Ultra $99.99 (5x Pro), and Ultra $199.99 (20x Pro). App limits refresh every 5 hours until a weekly limit. The page does not print a token cap for the app.

The next rates on the pages we checked

Claude's pricing page lists Sonnet 5.5 at $2 per million input tokens and $10 per million output tokens, the same sticker as Argon's intro price. Opus 5.5 is $4 and $20, which is the price Google says Argon moves to on December 1. Fable 5.1 is $10 and $50. US-only inference is 1.1x.

Cursor's model table lists Grok 4.7 at $2 per million input tokens, $0.50 cached, and $6 per million output tokens. Fast mode is $4, $1, and $12. Long context above 256k input tokens is billed at 2x, up to 500k. That table is Cursor's price for the model, not a second X Subscription.

Should you switch

If you are paying the API and the job fits in the intro window, Argon's $2 / $10 rate matches Sonnet 5.5's published input and output rates and undercuts the $6 Grok 4.7 output rate on Cursor's table. That comparison dies on December 1 unless Google changes the posted follow-on price. It says nothing about which model is better. We did not run them.

The subscription side of this, including who gets Argon inside an app, is on AI Model & Plan Prices.