Tech Brew Ride Home
Tech Brew Ride Home

Hey, Want Some Models?

September 2, 2026

AI Summary

5 min read

"Hey, Want Some Models?"

Anthropic, Google, and OpenAI all released or previewed major AI model updates on Wednesday, September 2, 2026, but the story is less about benchmark scores and more about how each company is wrestling with the same three problems: making agents capable enough to finish difficult work, economical enough to leave running for hours, and governable enough to trust with sensitive systems. Meanwhile, a federal judge spared Google's ad exchange from a forced sale, and the FBI opened an investigation into a breach exposing 153 million driver's licenses.


Anthropic's Fable 5.1: Enterprise economics and a new pricing trick

Anthropic released Fable 5.1—also known as Methos 5.1 in restricted versions for cybersecurity and life sciences organizations—and the headline numbers are strong. On Browserbase's hardest browser agent benchmark, Fable 5.1 completed 82% of tasks versus 74% for Opus V and 57% for Fable V. Investment firm Millennium told Anthropic the model traced an extremely rare software crash to a bug inside an external vendor library after the problem had resisted explanation for four to five years. Corporate expense management provider Ramp described an unattended 38-hour machine learning run in which the model reevaluated a previous result, launched six experiments, and returned with findings and proposed next steps.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

What you'll learn

  • 1 (00:34) **Episode Open & Ad** - Host Brian McCullough introduces the episode's top stories and reads a sponsored segment for Eight Sleep's Pod 5 smart mattress cover.
  • 2 (01:35) **Anthropic Launches Fable 5.1 / Methos 5.1** - Anthropic releases a new flagship model with two versions: a general-purpose Fable 5.1 and a restricted-access Methos 5.1 for cybersecurity and life sciences.
  • 3 (03:45) **Fable 5.1 Pricing & Economics** - The model's headline API rates are $10/M input and $50/M output, but the key change is a drastically reduced cached input price of $0.55/M, down from $1.00.
  • 4 (06:16) **Google Releases Gemini 3.8 Flash** - Google launches a new model just three weeks after 3.7 Flash, with an introductory price of $0.75/M input and $3.75/M output until December 31, 2026.
  • 5 (07:38) **OpenAI to Publicly Release Astra Model** - OpenAI plans a public release of its Astra model, which includes advanced coding and computer-operating capabilities, but will restrict cyber capabilities to testers.
  • 6 (11:23) **Mid-Roll Ads** - The episode includes two sponsored segments: AT&T Connected Car and Intuit Enterprise Suite.
  • 7 (12:12) **Judge Rules Google Does Not Have to Sell Ad Exchange** - A US federal judge decides Google must make its ad tech tools interoperable with rivals' tools, but avoids a forced sale of its ad exchange.

+ Full timestamped outline available in the app

Guests on this episode

Show Notes

Anthropic shipped Claude Fable and Mythos 5.1 with a 75% cache-price cut, Google countered with Gemini 3.8 Flash, OpenAI teased Astra's public release, a judge spared Google's ad exchange from a sale, and the FBI probed a 153M-license leak.

Tech Brew Ride Home