October 18, 2026
Heard in AI
Daily email

A story we follow

Google's Gemini 4 Argon launch

Google's 30 September 2026 announcement of Gemini 4 Argon, the first Gemini 4 model. Covers how capable and cost-effective it is compared with rivals, and the panel's explanations for its strengths.

A story page follows one specific event across podcast discussions, with an overview and a timeline of what changed. It updates when new episodes discuss the event, and you can get those updates by email or push. How our formats work

Updated

Overview

Google announced Gemini 4 Argon on 30 September 2026 as the first Gemini 4 model. Peter Diamandis presented it as Google's comeback after a promised Gemini 3.5 Pro never shipped, and he stressed long-horizon reasoning and much larger output limits. Alex Wissner-Gross judged it to be neither on the capability frontier nor on the cost-performance frontier. He said it puts Google back in the top three labs but not the top two, and that Google's highlighted benchmarks look somewhat cherry-picked. Richard Socher added that Google's business may not need superintelligence.

In an episode published on 17 October 2026, Nathan Labenz described a wide rollout on the preceding Tuesday and his own pre-release testing. He had Claude run the new model on the same context as earlier tasks and compare results side by side. He found it competitive and in the same class as the latest OpenAI and Anthropic models. Those models still won head-to-head comparisons somewhat more often, as judged by Claude and, in a limited way, by Labenz. He concluded that it puts Google back in the game. He cautioned that his sample was limited and tuned to the models he already uses. He also said the gain was not enough for him to swap models, because he now weighs security and predictability above small performance gains.

Recent coverage

Changes to our coverage published in the past seven days. These are site update dates; podcast discussion dates and sources are in each version.

No changes to this account were published in the past seven days.

What changed

Dates show when each podcast discussion was published.

  1. New information

    Nathan Labenz reported pre-release testing of Gemini 4, with Claude running side-by-side comparisons on his past tasks. He found it competitive and in the same class as the latest OpenAI and Anthropic models, though those still won head-to-head a bit more often. He said it puts Google back in the game. He caveated that his task sample was limited and tuned to his existing models, and that he would not switch because the step change was too small to justify extra security risk.

  2. New information

    The episode covers Google's Gemini 4 Argon announcement. It adds Alex's assessment that Argon is third among frontier labs and off the cost-performance frontier, with his speculation about why it excels at low hallucination.

Podcast discussions

Sources

  1. 01
  2. 02
  3. 03
  4. 04

Our coverage

Version history

  • Oct 18, 2026 · Version 2

    Nathan Labenz reported pre-release testing of Gemini 4, with Claude running side-by-side comparisons on his past tasks. He found it competitive and in the same class as the latest OpenAI and Anthropic models, though those still won head-to-head a bit more often. He said it puts Google back in the game. He caveated that his task sample was limited and tuned to his existing models, and that he would not switch because the step change was too small to justify extra security risk.

Your privacy

Necessary

Keep forms secure, remember your cookie choice and support editor sign-in.

Always on

With your permission, a random browser ID helps us count visits, return visits, article reading and short campaign labels. We keep these statistics on our own server for up to 180 days. Basic daily page counts may continue without this permission.

Refusing detailed analytics will not change your access to articles or the daily email.

Cookie details · Privacy