- Published on
Google Gemini 4 Argon closes the gap with OpenAI and Anthropic but doesn't take a clear lead
- Authors
- Name
- DeepHorus
Google has unveiled Gemini 4 Argon, its latest flagship AI model, which closes the gap with frontier rivals from OpenAI and Anthropic while beating some of them on key benchmarks. The model represents Google's first frontier release in more than seven months, following Gemini 3.1 Pro. It signals that the search giant is back among the top three AI labs, though Anthropic likely still holds the lead overall.
The release comes after a difficult and drawn-out development period during which the already-announced Gemini 3.5 frontier model was skipped entirely. Argon restores Google's competitive standing after months of falling behind rivals. It is initially rolling out to a group of trusted cyber defenders under the Fairwind program, with early feedback feeding into safety mechanisms.
According to independent tests by Artificial Analysis, Argon scores 53 points on its Intelligence Index at its highest reasoning level, tying OpenAI's GPT-6 Astra and Anthropic's Claude Fable 5.1. Anthropic's Claude Opus 5.5 leads with 58 points, followed by Claude Sonnet 5.5 at 56. Compared to Google's previous frontier model, Gemini 3.1 Pro Preview, Argon jumped 23 points.
A notable new feature is support for up to one million output tokens, up from 64,000, which Google calls an industry first. This allows the model to reason through complex problems in a single pass without timing out. A new Long Decode Continuation feature pauses and resumes long responses to further support extended reasoning.
Pricing begins at an introductory rate of $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 later. Cached inputs receive a 95 percent discount. Google has not announced a public release date, saying only that access will open as soon as possible.
