The AI Exhale: Astra Landed, Fable Led - The Week the Scorecard Broke (and the calculator in your pocket)
Two frontier models land inside 48 hours of each other. Claude Fable 5.1, then GPT-6 Astra. Both companies rate their own model ahead of the other, both on their own scorecard. The independent index puts Fable in front. Then OpenAI's president says welcome to the AGI era, and declines to define AGI. That tells you where we are. These models are commoditised and nobody is loyal to any of them. Anthropic ran away with it late last year, OpenAI went cheap in June, developers switched en masse. Speech and images to OpenAI, coding and finance work to Anthropic, and you bounce between them all week. Costs look similar on paper. The number that matters is how efficiently a model spends tokens to land a good answer. Two Ferrari builds side by side on socials, and the interesting column is what it cost to get there. As for AGI, outperforming humans at valuable work is something a calculator did to all of us in the 90s. This reads as a sales pitch for frontier leadership with a bit of panic under it, open source coming up from below and margins getting squeezed. Build on frontier, don't scale on it. This week didn't change that. Jez and Cam pull apart a scoreboard nobody should trust, and land on why choosing your own model will look ridiculous once routing gets good. Five minutes. One thing that matters. We ignore the rest, so you can too. Vibe CFO is Australian founded. An AI CFO for small and medium business, out of the box, governed and secure. Vibe in 5. vibecfo.ai