Google announces Gemini 4 Argon AI model, but you can't use it yet

Google announced Gemini 4 Argon, a new AI model it claims offers industry-leading performance in coding, knowledge work, and cybersecurity. The model is not yet available to the public. Google said engineers inside the company are already using it extensively.
According to Google, Argon used "fleet-wide telemetry data" to help save 300 TiB of memory across its data centers. Argon agents have also worked to migrate C/C++ codebases to Rust across Google, including thousands of lines in the core re2 and libgav1 libraries and more than 800,000 lines in the Fuchsia OS Zircon kernel.
Google cited benchmarks for its claims. On the software engineering DeepSWE v1.1 benchmark, Gemini 4 Argon scored 77.9 percent, higher than GPT-6 Astra, Fable 5.1, and Opus 5.5. Google also pointed to Argon's industry-leading score on the economic analysis Vals Index test and promised similar power across a range of long-horizon tasks.
The model is still in limited testing, but Google announced API pricing: for a limited time, $2 per million input tokens and $10 per million output tokens, with cached input tokens discounted 95 percent. Google confirmed Gemini 4 Argon will support an output limit of 1 million tokens, up from 64,000 tokens in previous Gemini models, which Google says allows users to complete more daunting tasks in a single step.
Google said the main focus for Gemini 4 right now is cyberdefense, and that models of this scale call for a phased release, so it is starting with a small group of trusted testers. Partners in the company's Fairwind Program can get access to the model's cybersecurity defense capabilities. Wiz is already using Argon and used it to uncover a critical vulnerability that could expose personal information in a system used at hospitals around the world, according to Google, which claims other frontier models missed this flaw but did not provide specifics.
Citing concern about model misalignment after several high-profile hacking incidents over the summer, Google claims it designed Argon with systems that monitor the model's chain-of-thought and can stop it if the model steps out of bounds. Google said Gemini 4 Argon will eventually be available to enterprise and consumer customers but gave no specific timeline, saying general availability will start with paid API users and Google AI Ultra subscribers after the cybersecurity tests.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
So much for Gemini 3.5 Pro.