Gemini 4 Argon (blog.google)

Article summary

Google has announced Gemini 4 Argon, a new AI model that delivers advanced reasoning in complex workflows, including software engineering, enterprise knowledge work, and cybersecurity defense. The model features a 1 million token limit and is currently being rolled out to trusted cyber defenders through the Fairwind Program. Google is prioritizing safety and rigorous testing before a wider release to the public. The model will be available at an introductory price of $2 per million input tokens and $10 per million output tokens.

Main themes

• AI model releases
• Access and safety
• Pricing and competition
• Model flaws and biases
• Cybersecurity and defense
• AI risks and consequences

What commenters say

• Some commenters are skeptical of Google's decision to limit access to the new model, citing concerns about the company's ability to release functional models.
• Others believe that the model's capabilities are too powerful to be released to the public without proper safeguards.
• The pricing of the new model is seen as competitive with other AI models on the market, but some commenters are concerned about the potential for increased costs after the introductory period.
• There are also concerns about the model's potential flaws and biases, with some commenters citing past experiences with Gemini models as evidence.
• Some commenters believe that Google is following a standard playbook for AI model releases, including limiting access and creating hype around the model's capabilities.
• Others are worried about the potential risks and consequences of releasing advanced AI models, including the possibility of 'paperclipping' or other catastrophic outcomes.