Gemini 4 Argon is live worldwide at promo pricing — but it sits #3, not #1, on the latest Vals Index

Asked (summary):

"gemini 4 argon" — interpreted as a concise, current research profile of Google Gemini 4 Argon as of 2026-10-26.

This profile draws on 22 source-backed rows across 7 hosts — official Google blog, docs and pricing pages plus the independent Vals AI leaderboard — covering the 2026-09-30 announcement through the scheduled 2026-12-01 price change. URL query variants duplicate several documentation pages; they are consolidated to 10 canonical pages below.

Launch to current: the Argon timeline

Hover a marker for detail. Blue: past events. Black: today. Amber: scheduled pricing change.

The unified budget is 1,048,576 tokens for input, reasoning and output combined; long-decode output can reach the same ceiling minus input tokens — not a guaranteed 1M-token response. ai.google.dev

Promotional API pricing through 2026-11-30 is $2 input, $0.10 cached input and $10 output/reasoning per million tokens; from 2026-12-01 it doubles to $4, $0.20 and $20. cloud.google.com

Google's launch post said Argon ranked #1 on the Vals Index, but the latest leaderboard evidence (2026-10-19) places it #3 at 69.24%, behind Claude Sonnet 5.5 (70.85%) and Claude Opus 5.5 (70.18%); positions change over time. vals.ai

Ordinary Argon refuses harmful cyber and CBRN requests; a cyber-permissive variant without cyber guardrails is restricted to approved defenders through Google's Fairwind Program. blog.google

Model facts

API model IDgemini-4-argon

Unified token budget1,048,576 (input + reasoning + output)

Max outputup to 1,048,576 minus input tokens

Inputstext, images, video, audio

Outputtext

Announced2026-09-30

Globally available2026-10-13

Vals Index (2026-10-19)#3 at 69.24%

Pricing per million tokens

Input — now (through 2026-11-30)$2.00

Cached input — now$0.10

Output / reasoning — now$10.00

Input — from 2026-12-01$4.00

Cached input — from 2026-12-01$0.20

Output / reasoning — from 2026-12-01$20.00

Access routes

Developers and enterprises: paid Gemini API in AI Studio, Android Studio and the Gemini platform
Google AI Pro and Ultra subscribers: Gemini app and AI Mode in Search
Google AI Ultra subscribers: Google Antigravity
Scientists and researchers: can apply for 6 months of free access to Antigravity's highest usage tier, subject to eligibility

Best-fit workloads

Long-horizon, multi-step reasoning and deep work
Real-world software engineering and coding
Enterprise knowledge work: finance and legal
Cybersecurity defense (guardrailed for general users)
Science, visual understanding and writing

Caveats

The 1M figure is a unified token budget, not a guaranteed 1M-token response: streaming/non-streaming defaults and input length limit output in practice
Full cyber capability without guardrails is reserved for trusted defenders via the Fairwind Program and Google's internal teams
The launch-era "#1 on the Vals Index" claim is not the current ranking; the 2026-10-19 leaderboard shows #3
Current prices are promotional and end 2026-11-30; budget for a 2× increase from 2026-12-01

Source-linked evidence

Page	What it establishes	Rows	Source

Research profile of Google Gemini 4 Argon as of 2026-10-26. 22 source-backed rows across 7 hosts (Google blog, docs, pricing pages and independent Vals AI), consolidated to 10 canonical pages; URL query variants of the same page are counted as one source, not independent corroboration. Prices are USD per million tokens; token figures are the model's unified input + reasoning + output budget. Full positioning and availability text shortened for space.

This report was generated automatically by Keenable SELECT at a user's request, from publicly available web sources linked herein. Keenable does not review, verify, or endorse its contents and makes no representation as to accuracy, completeness, or timeliness; AI-based extraction may contain errors. Nothing in this report is investment, legal, financial, or other professional advice. All trademarks and referenced content remain the property of their respective owners; no affiliation or endorsement is implied. To report an error, rights concern, or request removal: legal@keenable.ai.