Google AI Updates: October 1, 2026

1. Google Announced Gemini 4 Argon, Starting With Vetted Cyber Defenders

Google. Google DeepMind introduced Gemini 4 Argon, a frontier model aimed at software engineering, enterprise knowledge work in legal and finance, and cybersecurity defense, with up to 1 million output tokens (up from 64K). Google reported 77.9 percent on DeepSWE v1.1, 51.3 percent on AutomationBench, 91.7 percent on LVBench, and 68 percent on CWE-bench v1, with introductory API pricing of $2 per million input tokens and $10 per million output tokens (rising later to $4 and $20) and a 95 percent discount on cached input. Access is initially limited to trusted cyber defenders through the Fairwind Program, with paid API customers and Google AI Ultra subscribers named as the next phase but no date given, so most developers cannot call the model yet. Source

2. DeepMind Released SynthID Bio to Watermark AI-Designed Proteins

Google. SynthID Bio embeds detectable watermarks into AI-generated protein sequences by steering amino acid choices, and into predicted 3D structures by fine-tuning AlphaFold 3’s diffusion network to shift atomic coordinates. In wet-lab tests on designs against VEGF-A, the SARS-CoV-2 spike RBD, and PD-L1, watermarked proteins matched the hit rates and binding affinity of unwatermarked ones, with near-perfect detectability. DeepMind published the methods paper and is open-sourcing code, data, and weights, giving biodesign tool builders a provenance mechanism to evaluate for biosecurity screening. Source