✳ the wire · analysis

Gemini 4 Argon ships at $2 / $10 until December 1, then the price doubles

Gemini 4confirmedby ArtificialWatch

Get launch alerts like this, free →

Gemini 4 Argon ships at $2 / $10 until December 1, then the price doubles
Source imagery · verified against a primary source

Gemini 4 Argon is generally available. Google's Gemini API changelog lists gemini-4-argon as GA on October 13, "our most capable model built for complex coding, agentic, math, and science tasks", and it is in AI Studio, Vertex AI, the Gemini app and Antigravity. Our sweep fired at 15:45 UTC.

THE PRICE, AND THE DATE IT DOUBLES

$2 per million input tokens and $10 per million output through November 30, 2026. From December 1 it is $4 and $20. Cached input is $0.10 per million, then $0.20, and only implicit caching is supported. Batch and Flex run at half price; Priority costs $3.60 and $18 for now. Argon is not on the Free tier or the Google AI Pro plan.

SPECS

A 1,048,576-token context window that input and output share. Replies default to 65,536 tokens and can run to about 1M with what Google calls long decode. Thinking comes in low, medium and high. It reads text, images, video, audio and PDFs and writes text. The model card puts its knowledge cutoff at March 2026.

GOOGLE'S OWN FIGURES

Google's table compares Argon with GPT-6 Astra, Claude Fable 5.1 and Claude Opus 5.5. Against Opus 5.5 it is ahead on 14 of the 18 rows where both have a score, including DeepSWE v1.1 (77.9% to 74.2%), the Vals Index (68.9% to 67.0%) and long-context GraphWalks at 256K to 1M tokens (84.2% to 66.8%). It is behind on four: FrontierSWE v2 (55.0% to 62.3%), Terminal-Bench 4.0 (57.4% to 66.4%), PostTrainBench and Terminal-Bench Science. These are Google's numbers.

THE FIRST INDEPENDENT SCORES

Artificial Analysis scores Gemini 4 Argon 52.6 on its Intelligence Index at high, Google's top setting, when we read it on launch day. That is level with GPT-6 Astra at max (52.7) and behind Claude Opus 5.5 (57.6), Claude Sonnet 5.5 (56.0), Muse Wonder Full (55.8) and Claude Fable 5.1 (53.4). Vals has it first on the Vals Index at 68.9%. Across the three boards that already score it, our composite on /benchmarks puts it fifth.

WHAT WE DO NOT KNOW

Whether the December price doubling sticks or slides. And whether Barium and Carbon, the newer Gemini 4 checkpoints Business Insider reported on October 9, ship as separate models. Both names are on our radar under the next Gemini 4.

Source: Gemini API changelog, pricing and model docs, read directly · Google DeepMind model card · Artificial Analysis, read directly · our own sweep ↗ · Gemini 4 tracker · the bench index

sweeping every 60 seconds

Hear the minute Google’s next model drops — not the minute we write it up.

Claude Opus 5 went live at 16:51 UTC. The alert was in subscribers' inboxes at 16:52.

  • Free forever
  • No card
  • Unsubscribe in one click

← back to the wire