Google's Gemini 4 Argon rollout

Google DeepMind announced Gemini 4 Argon on September 30, 2026, releasing it initially to trusted testers through a program called Fairwind, which covers trusted testers, internal Google staff, and vetted cybersecurity defenders. API customers and AI Ultra subscribers are the next access tier before broader availability. The model is not yet publicly accessible.

Google's benchmark claims place Argon at 77.9% on DeepSWE v1.1, 91.7% on LVBench, 51.3% on AutomationBench, and 68% on CWE-bench v1. Across 19 benchmarks, Argon leads outright on 13, but GPT-6 Astra leads on several software, science, and computer use benchmarks, and Claude Opus leads Terminal-Bench and PostTrainBench. On finance and legal agent benchmarks specifically, Argon scores 65.4% versus GPT-6 Astra's 53.5%, and 19.6% versus 5.4%. Vals AI, an independent evaluation platform, ranks Argon first out of 41 models with 68.90% accuracy on its index. AlphaCorp AI notes that every figure published on September 30 comes from Google, has been replicated by no outside lab, and arrives without a technical or safety report.

The output token limit is 1 million, which one report describes as nearly 8x the 128K cap on GPT-6 Astra, Claude Opus 5.5, and Fable 5.1. On prompt injection resistance, Google reports a 0.7% attack success rate after 15 attempts on the Gray Swan IPI benchmark, compared to 52.7% for Kimi K3.

Pricing is listed at $2 per million input tokens and $10 per million output tokens, with cached input at 95% off; one source describes this as exactly half the cost of Claude Opus 5.5. Vals AI lists the price at $4 per million input and $20 per million output, diverging from other sources.

Internally, Argon agents freed over 300 TiB of data-center memory with total estimated savings of 500 TiB to 1 PiB, and are migrating C/C++ codebases to Rust, including over 800,000 lines in the Fuchsia OS Zircon kernel and work on re2 and libgav1. On libgav1, Argon replaced 32,000 lines of SIMD code to make the Rust port 2.7x faster.

For cybersecurity, Google stated: 'For trusted defenders and our own internal teams at Google, we'll be releasing Argon without cyber guardrails so they can leverage its full frontier-level cybersecurity defense capabilities.' Safety measures include chain-of-thought inspection for misalignment monitoring and hardened sandboxed environments. Google has engaged with the U.S. government's voluntary pre-release model access process as part of the phased rollout.