Gemini 4 Argon Released: Analyzing Benchmarks and Model Card

On October 13, Google officially began the public rollout of its new language model, Gemini 4 Argon, which was announced in late September. The model is now available via a paid API, and users with Pro and Ultra subscriptions have gained access within the Gemini app. Alongside the release, the company published detailed documentation revealing technical specifications, benchmark results, and model architecture details. A key aspect of the update includes data on risks identified during internal testing, as well as information regarding the duration of promotional pricing for developers. Google emphasizes that Gemini 4 Argon demonstrates significant progress in logical reasoning and complex context processing compared to previous versions. Analysts note that this release is a major step in Google's competitive push in the generative AI market, offering higher performance for both enterprise and individual users.