GPT 5.5 xhigh is smarter than Fable but Fable like Opus 4.8 as well is faster and seems more “agentic”. It’s easy to test this. Build a fairly complex software with Claude(opus or Fable).
Review the commits with both Claude and GPT 5.5 Xhigh. You can see that Fable is still sloppy(er) compared to GPT. You can test it the other way around as well(drive the dev with GPT and review with GPT and Claude). You get the same result
Claude has an edge though and that’s on building more beautiful user interfaces.
Review the commits with both Claude and GPT 5.5 Xhigh. You can see that Fable is still sloppy(er) compared to GPT. You can test it the other way around as well(drive the dev with GPT and review with GPT and Claude). You get the same result Claude has an edge though and that’s on building more beautiful user interfaces.