Google opens Gemini 4 Argon globally, pitching one model for robots and enzyme design
Google DeepMind made Gemini 4 Argon globally available on October 13 to eligible paying subscribers, developers and enterprises, after its September 30 debut was limited to selected cybersecurity defenders. Access routes differ: the Gemini app and AI Mode in Search require a Google AI Pro or Ultra subscription, the API and Google AI Studio require paid tiers, and Antigravity requires Ultra, while free Gemini users are excluded; introductory API pricing is $2 per million input tokens and $10 per million output tokens through November 30, doubling on December 1, with a 95% discount on eligible implicit-cache hits. Google also showed Argon working in science and robotics: researchers used it to organize data, tune models and set screening rules within hours, producing five serine esterase designs that showed catalytic activity in lab tests, while in robot demos it built a MuJoCo simulation from a single photo, generated more than 4,500 training episodes and lifted real-robot task success from 56% to 75% while cutting task time from 35 to 24 seconds.
General frontier models are moving from chat into robotics and lab workflows, though real-world results and self-reported benchmarks still need independent verification.