Theo Reveals GPT-5.6 Capabilities For Agent Orchestration And iOS Dev
Theo Browne says OpenAI's new model fixes many of his GPT-5.5 complaints and works especially well for long-running agent tasks.
4467.1K181587430.2K
“Apparently I'm allowed to talk about GPT-5.6 now? It's a damn good model. Not quite as "smart" as Fable, but it is incredibly capable. Fixed all the problems I had with GPT-5.5.”
Theo Browne@theoTECH#1329Story Brief
In a post on X, Theo Browne said OpenAI's GPT-5.6 is "a damn good model" that fixed the problems he had with GPT-5.5, especially around long-running agent work, subagent orchestration and iOS development. Browne said the `gpt-5.6-sol` variant will become his default for many tasks, while adding that it still has rough edges. In a separate post on X, he said OpenAI asked early testers to hold formal writeups until Thursday while allowing public discussion in the meantime.
“I basically stopped using Hermes when I had to move back to 5.5. So relieved to have it back. Night and day difference.”
Theo Browne@theo
