Google announced Gemini 4 Argon on September 30, 2026, as its new frontier large language model with claimed state-of-the-art results on benchmarks for long-horizon software engineering, enterprise knowledge work, and cybersecurity defense, including a 77.9% score on DeepSWE v1.1 that edges out recent Claude Opus 5.5 and GPT-6 Astra releases. Initial access is limited to internal teams and vetted participants in the Fairwind Program for cyber defenders, with Google citing the need for phased rollout, guardrail iteration, and U.S. government pre-release review before expanding to paid API users, Google AI Ultra subscribers, developers, and enterprises. The model features a 1-million-token output limit and introductory pricing of $2 per million input tokens and $10 per million output tokens. Traders are watching for the timing of broader availability, which Google describes as “as soon as possible” without a firm date, alongside any competitive responses from OpenAI or Anthropic.