major price cuts today:
*80% drop for GPT-5.6 Luna, now $0.20 per million input tokens and $1.20 per million output
*20% drop for GPT-5.6 Terra, to $2/$12
*GPT-5.6 Sol gets Fast mode in the API, up to 2.5x the speed for 2x the price, same intelligence
Most AI unicorns have published 0 scientific papers. The few that do make up 0.1% of all AI research. And within that, OpenAI alone owns ~40% of the citations.
https://t.co/qCzp8JE43s
@bryan_johnson@grok what do you think? I feel there are some people pulling him down thats demotivating him or something? He should keep at it right? When people like him push things too far thats when something good happens for the world!
Training the "best" model has stopped being a moat. The open/closed gap is now 4-7 months - Kimi K3 is near Opus-level at a fifth of the price.
Maybe that's why companies like Vercel, Ramp, OpenRouter, Cursor, etc. are all building routers and not models. They're also getting valued like crazy.
Eventually whoever has more access to compute - the ability to serve billions and trillions of tokens - may just end up making massive revenue on competitive, razor-thin margins.
@zeeg That’s not bad right? You’re paying 100pc more for 50pc more productivity. Going forward assuming cost per task goes down (more intelligence, less token price, combo of both) - this is just transitional. Not something “can’t scale at all”