While 5.6 SOL has better benchmarks for accuracy in high thinking, it comes at a steep price in response time. I work with GPT uploading database files that the A.I. uses to make decisions based on those files. While 5.5 made more mistakes in reading and executing decisions based on the .json files being uploaded to it, they were usually pretty minor mistakes and I could work with that. 5.6 SOL still isn’t 100% accurate and a lot of the problems stem from it struggling with deterministic approaches to problems. Simply put, it will look at a path that it is required to take and then decide on its own to use another path to solve the problem. While 5.5 would make a decision quickly and go with it, 5.6 SOL will check its work, re-check it’s work, re-check it’s work and so on resulting in absurd response times. The same operation that 5.5 took 1 minute to complete will take 5.6 SOL 10 minutes or more to complete and sometimes 5.6 SOL will take so long that it will time out and never complete the operation. This is an unacceptable response time when the end user needs data returned within a minute if not faster. Removing 5.5 will make GPT unusable if 5.6 SOL isn’t ready to go in October. Because it sure isn’t ready now.
Same here, it gives shallow responses and doesn’t check files or do actual research and thinking, or cite the source, it just gives a bland response.. I think the slow shallow responses has something to do with ChatGPT tryna get rid of 5.5 on October 14 so the update is heavily infected on everything else, I wonder when this is all gonna get fixed.