The most prestigious law school admissions discussion board in the world.
|
 |
 
|
Codex sol ultra is significantly better than Fable 5
| The Penis | 07/17/26 |
Poast new message in this thread
Date: July 17th, 2026 4:58 PM Author: The Penis Fable 5 is better than gpt 5.6 if you are just using the prompt-response chat interface. But for long running agentive workflows codex ultra destroys it. I remember gpt 5.5 running agent workflows on codex actually beat Fable 5 on an agent benchmark and it was a huge upset. I can see why even gpt 5.5 beat it, because sol ultra is wild and also seems like Open AI is trying to move away from the chat interface model more towards agents (http://www.autoadmit.com/thread.php?thread_id=5883629&forum_id=2Firm#50006999) |