Google announces Gemini 4 Argon

Google describes Argon as a model built to sustain deep reasoning across complex, long-horizon workflows.[1] It says the model delivers frontier performance for software engineering, enterprise knowledge work and cybersecurity defense.[1] Google DeepMind describes Argon as a large-scale sparse mixture-of-experts model and says it is not a modification or fine-tune of a prior model.[4] For teams assessing it, those descriptions set out the intended use cases and model design, but the announcement’s claims should be distinguished from independently verified performance results.

Google says Argon is globally available to developers and enterprises through the API in AI Studio and the Gemini platform.[2] Google AI Ultra subscribers can access it in Google Antigravity, while Google AI Ultra and Pro subscribers can use it in the Gemini app and AI Mode in Search.[2] Google also describes Argon as available to developers, enterprises, scientists and consumers.[2] Cybersecurity access is a separate, staged rollout: Google is providing the model to trusted cyber defenders through its Fairwind Program and says it will gather early-tester feedback and iterate on guardrails before wider availability.[1]

Google’s developer documentation lists introductory prices of $2 per million input tokens and $10 per million output tokens through Nov 30, 2026; it lists later prices of $4 and $20 per million, respectively.[3] Google says cached input tokens receive a 95% discount from the input-token price.[1] Argon supports a unified 1,048,576-token context window shared across input, thinking and output, and can generate a single response of up to 1 million tokens minus the input.[3] These figures give teams concrete parameters for testing cost and long-document workflows.

Google says thousands of Googlers highlighted Argon’s strengths in specialized coding, deeper research and writing quality.[1] Google says Argon ranks first on the Vals Index, which measures performance across economic sectors including finance, coding, legal and tax work.[2] In Antigravity, Google says Argon helped teams increase agentic code submissions by more than 35% and reduce code rollbacks by nearly a third.[2] These are company-reported results.