Google's new frontier AI model Gemini 4 Argon goes to cybersecurity defenders first - SiliconANGLE

Google has begun rolling out Gemini 4 Argon, described as a frontier AI model, to vetted cybersecurity defenders. The rollout was reported by SiliconANGLE. According to the report, the model beats rival models from Anthropic and OpenAI on most of Google's benchmarks. The article's headline states that the model "goes to cybersecurity defenders first." The report does not identify which benchmarks were used, how they were selected, or the size of the margins involved.

Google's Koray Kavukcuoglu wrote that releasing capabilities at this level "requires a phased approach." The company says it is taking part in the U.S. government's voluntary process for pre-release model access. Google also says it is still hardening safeguards against misuse. The report describes that access process as voluntary for the participating company. The excerpt does not state how many defenders have received access, which organisations they work for, or when any wider availability might follow.

The report frames the release as going to cybersecurity defenders first, rather than to general availability. It places Gemini 4 Argon alongside rival frontier systems from Anthropic and OpenAI, which the model is said to outperform on most of Google's own benchmarks. Google's stated participation in the U.S. government's voluntary pre-release access process is presented as accompanying the rollout, alongside continued work on safeguards against misuse. Beyond the benchmark comparison and the defender rollout, the excerpt provides no further technical detail about the model.