Developer wants to see GPT-6 Sol results on a new benchmark

iannuttall says a new benchmark feels closer to everyday model use and names GPT-6 Sol as the result he wants to see, an expectation rather than a test.

Ian Nuttall

@iannuttallArchived 2026.09.24Originally posted 2026.09.23

Ian Nuttall @iannuttallOriginally posted 2026.09.23

Open original on X

I like this benchmark. It feels more like how these models are used in day-to-day work. I’ll be interested to see how GPT-6 Sol fares.

This post quotes another post: x.com/pawelhuryn/status/2102479269335101504

x.com/iannuttall/status/2102503635477557493Archived 2026.09.24 10:10

Open original on X