What changed
Google launched Gemini 4 Argon on October 13, its first major frontier-model release since Gemini 3 in late 2025.
Argon is rolling out to Pro and Ultra subscribers in the Gemini app and Search AI mode, plus Ultra subscribers using Antigravity. Google says it beats GPT-6 Astra, Claude Fable 5.1 and Claude Opus 5.5 across coding, cybersecurity, science, finance and law benchmarks.
The introductory price is $2 per million input tokens and $10 per million output tokens through the end of November. Those prices then double. Availability for Plus subscribers and free users has not been specified.
Why it matters
This is Google trying to turn a late arrival into a pricing advantage. Argon enters the frontier-model market at roughly one-fifth the per-token cost of the cited OpenAI and Anthropic models during its introductory period. That could make experimentation cheaper, but token prices alone do not determine the final bill: different models may use different numbers of tokens for the same task.
Google is also pointing to internal work rather than chatbot polish alone. It says Argon has helped find and patch cybersecurity vulnerabilities, halve the computing power used for YouTube’s comments section and debug the drone-physics simulator for Wing.
Those claims suggest the real test will happen inside complex systems, where reliability and resource use matter more than a strong benchmark score. Google first gave the model to a limited group of trusted cyber defenders to test potential risks, but independent results and real-world comparisons with rival models remain unconfirmed.
Comments
No comments yet.