OpenAI’s “generational leap” with GPT-6 Astra

PLUS: Where the future of content is headed

Zach Mink

September 4, 2026

Good morning, AI enthusiasts, and welcome to our 3,363 new readers. The long-awaited GPT-6 Astra has finally arrived. OpenAI calls it "a new generation of intelligence," with president Greg Brockman going further with "Welcome to the AGI era."

So has AI's loftiest, least-defined bar actually been cleared? Brockman says he'll leave that to the reader (once you can actually use it), though the early benchmarks may already be making the case for him.

In today’s AI rundown:

OpenAI’s “generational leap” with GPT-6 Astra
Rowan’s Corner: Where the future of content is headed
Use the loop method for better ChatGPT results
Google's WeatherNext 3 sharpens the forecast 5x

LATEST DEVELOPMENTS

OPENAI

📶 OpenAI’s “generational leap” with GPT-6 Astra

Image source: OpenAI

The Rundown: OpenAI just introduced the long-awaited GPT-6 Astra, calling it the “most intelligent and aligned model in the world,” with benchmarks setting new highs across domains like science, math, computer use, coding, cybersecurity, and more.

The details:

Astra hit 99.9% on ARC-AGI-3, up from GPT-5.6 Sol's 7.8%, with other notable scores on FrontierMath T4 (98%), and 100% on ExploitBench for cybersecurity.
The model lands at 61 on AA’s Intelligence Index, behind Fable 5.1, 5, Opus 5, and Meta’s Muse Spark 1.3, despite highs across a variety of other tests.
Astra is priced at $10/$50 per million tokens in the API, roughly 2.5x more expensive than GPT-5.6 Sol but with more efficient token usage per task.
A small group of organizations gets it first, with paid ChatGPT plans and the API following within days and an Astra Pro tier for Pro plans and up.
When asked if Astra qualifies as artificial general intelligence, OpenAI President Greg Brockman said: “For me personally, I do think we're there."

Why it matters: It’s a letdown to see the biggest release of the year not immediately available at launch, but Sam Altman said: “It should be quick.” Astra’s benchmarks are powerful, but the real test comes when users get access — and with Fable 5.1 here at the same price, the frontier now has a pretty clear head-to-head to compare against.

TOGETHER WITH GOOGLE FOR STARTUPS

⚒️ Build a production-ready, multi-agent co-founder suite

The Rundown: Startup School: Agent Builder is a 4-part live series designed specifically for founders and lead engineers. Kicking off on Sept. 15, you’ll master the Gemini Enterprise Agent Platform and Google Antigravity 2.0 to transition your AI from rapid prototyping to enterprise-grade production.

In this intensive series, you will:

Code agent state transitions in Python with Google's ADK, with zero vendor lock-in
Build long-term agent memory with multimodal RAG and session persistence
Deploy secure, rate-limit-free agents at enterprise scale on Cloud Run

ROWAN’S CORNER

🔮 Where the future of content is headed

Image source: The Rundown’s YouTube interview with Sundar

Rowan: I recently had the opportunity to sit down with Google CEO Sundar Pichai for an exclusive interview, and one line has stuck with me since:

“I'm very confident three years from now you will look back at what we have today and realize it was pretty primitive. It's like looking at a flip phone.”

He was talking about content. His bet is that within three years, “any content can go from any modality to any modality.”

For example, YouTube clips could become blogs in one click, and newsletters could become podcasts without recording. Essentially, consumers would choose the format, and creators just supply the ideas, brand, and personality.

Net-net, I think this is a good thing, since it kills the "more content" strategy. When anyone can produce everything, volume stops being an advantage, so the creators and media brands who win will focus on quality. If the content is good, people don't care how it was made.

But I’m curious: if you had the option to turn this daily newsletter into a podcast, would you? Let me know in the poll below.

AI TRAINING

🔂 Use the loop method for better ChatGPT results

The Rundown: In this guide, you will learn how to improve any repeated ChatGPT workflow with the review-based "Loop Method.” We created it to improve our video editing and image generation skills, but you can try it on anything!

Step-by-step:

Open ChatGPT Work and find a workflow that needs improvement: a skill, project folder, or process that works inconsistently
Prompt: “Improve this workflow in 3 loops. Have a panel of sub-agents adversarially review each loop. Done when [definition of done]”
Use /goal to assign ChatGPT a deliverable you want the workflow to build. Let it run three loops, taking anywhere from 15 minutes to an hour
Test the improved workflow. If it still needs work, tell Codex to codify ambiguous work into scripts

Pro tip: Turn this self-improvement process into a reusable skill, then run it whenever a workflow becomes inconsistent.

PRESENTED BY GARTNER

📊 Everyone’s adopting. Few are winning.

The Rundown: By 2030, 0% of IT work happens without AI. Success isn’t determined by whether you adopt — it's how you adopt. Gartner helps leaders pinpoint what’s actually working, so you can act on what really matters for your organization.

With Gartner AI insights, you can:

Access 4,000+ proven AI use cases
Get objective, analyst-led guidance
Stay ahead

Explore Gartner AI Hub.

GOOGLE

🌦️ Google's WeatherNext 3 sharpens the forecast 5x

Image source: Google

The Rundown: Google DeepMind introduced WeatherNext 3, its new weather model that uses live satellite images to refresh its forecasts every hour, taking the top spot on scoreboards against both rival models and the U.S. and European weather services.

The details:

WeatherNext 3’s improvement is about 5x in detail for temperature down to 5 KM, enough to pick out individual valleys and coastlines.
Older AI forecasters learned from physics runs that came roughly 6 hours late, with WeatherNext adding in live feeds and ground stations for quicker turnaround.
Rain is typically AI forecasting’s weak spot, and Google said errors dropped up to 60% vs. satellite data, with day-ahead rain forecasts now up to 50% better.
Google is integrating WeatherNext 3 across its suite of apps including Search, Maps, Gemini, and Earth.

Why it matters: How a forecast ends up in your weather app isn't something most people think about, but the model underneath getting faster, more accurate, and available worldwide is a win nearly everyone can get behind. It's also another reminder that some of the biggest AI wins may be the ones nobody thinks of as AI at all.

QUICK HITS

COMMUNITY AI WORKFLOW OF THE DAY

▸ Sam built a good-news-only feed to replace doomscrolling

Today’s workflow comes from reader Sam:

“I read the news every day, but it had become increasingly depressing and was making me miserable. What bothered me most was that it also felt inaccurate: although bad things are happening, there have been many positive developments over the past five years that rarely receive sustained coverage.

I wanted a way to get only positive news stories in my feed each day. Keyword filters did not work: "record" and "breakthrough" also appear in stories about record wildfire seasons, while "war" can appear in "war ends." So I trained a basic artificial version of my personality using a series of prompts about what I consider positive in the world. I connected it to Mistral through OpenRouter and gave it access to public RSS feeds from news sites I already trusted.

This eventually became Rally News, which I published on Google Play. The app surfaces positive stories from more than 20 news sites in an endless scroll, giving me an alternative to my uncomfortable TikTok addiction. I built the app without coding experience for about $25 per month.”

See Sam’s full workflow Visit The Rundown University. How do you use AI? Tell us for a chance to be featured.

🛠️ Trending AI Tools

🎨 Gridset — Turn Claude docs, Notion pages, or PDFs into on-brand content using proven design systems built by designers*
🚀 Claude Fable 5.1 - Anthropic's new flagship, top-rated model
🗣️ Realtime-TTS-2 - Inworld's top-ranked voice model, now generally available
🌎 Atlas - World Labs’ new SOTA world model in early access

*Sponsored Listing

📰 Everything else in AI today

Nvidia announced the acquisition of Hugging Face for $12.9B, with CEO Jensen Huang saying the platform will remain open to every cloud and chip.

Axiom Math published a paper breaking a decade-old record by proving prime numbers never stop appearing within 212 of each other, only for OpenAI's GPT-6 Astra to shrink that gap even further to 186 on the same day.

Sen. Bernie Sanders teamed up with Rep. Greg Casar on a new bill to outlaw superintelligent AI outright, with penalties modeled on laws for nuclear weapons.

Mira Murati’s Thinking Machines Labs is reportedly in talks for a new funding round that would value the company at $40B, according to The Information.

Widespread outages occurred on Thursday across the major AI providers, with OpenAI, Anthropic, xAI, and Google all facing stability issues throughout the day.