Reported launch
Google's Gemini 4 Argon frontier model

- Google's new Gemini 4 Argon model excels on benchmarks but faces skepticism from employees about its practical performance.
Bloomberg reported that Gemini 4 performed well on benchmarks but less well when employees used it, especially in coding.
- Two people familiar with Gemini 4 told
Bloomberg the model appears affected by "benchmaxxing," optimizing for tests rather than useful work.
Yahoo Finance reports that Google is offering Gemini 4 Argon at introductory rates of $2 per million input tokens and $10 per million output tokens.
In this conversation · 1
30 SepGoogle Gemini 4 Argon available to trusted few
CSOonline,
EasternEye,
Security Affairs, and
The Decoder report that Google has made its Gemini 4 Argon AI model available to a limited group of trusted testers through the Fairwind Program. The model features a 1 million-token output limit and is priced at $2 per million input tokens and $10 per million output tokens initially. The outlets note that Argon closes the gap with rivals OpenAI and Anthropic but does not take a clear lead.
- Argon is initially available only to trusted cyber defenders through Google's Fairwind Program.
CSO
- The model's output limit has been increased from 64,000 to 1 million tokens.
The Decoder
- Introductory pricing is set at $2 per million input tokens and $10 per million output tokens, with regular pricing rising to $4 and $20 later.
The Decoder
- Argon scored 68% on the CWE-bench v1 vulnerability remediation benchmark, tying for first place with Grok 4.7, GPT-6 Astra, and Claude Opus 5.5.
CSO
- Argon is initially available only to trusted cyber defenders through Google's Fairwind Program.
Source details
- Latest report
- 30 Sep 2026 20:40 UTC
- First seen here
- 30 Sep 2026 21:25 UTC
- Outlets
- 14