Google announces Gemini 4 Argon AI model, but you can't use it yet
Google promised Gemini 3.5 Pro in June, but it spent the summer trotting out smaller Flash models. Now, Google is ready to take on the frontier again with Gemini 4 Argon. The company claims this new AI offers industry-leading performance in coding, knowledge work, and cybersecurity, but you aren’t allowed to use it yet.
TL;DR
- Google's new AI model, Gemini 4 Argon, is claimed to offer industry-leading performance in coding, knowledge work, and cybersecurity.
- Currently, Gemini 4 Argon is in limited testing, with internal engineers using it extensively and partners in the Fairwind Program gaining access to its cybersecurity features.
- Argon has reportedly saved 300 TiB of memory in Google's data centers and is assisting in migrating C/C++ codebases to Rust.
- The model shows strong performance on benchmarks like DeepSWE v1.1 and the Vals Index test.
- Gemini 4 Argon will support an output limit of 1 million tokens, significantly higher than previous models.
- Cyberdefense is a primary focus, with early access partners like Wiz already using it to uncover vulnerabilities.
- Google has implemented systems to monitor Argon's reasoning and ensure transparency.
- General availability is planned for paid API users and Google AI Ultra subscribers after cybersecurity tests are completed.