ChatGPT (o3 & o4-mini Models) Use of Multiple Placeholders

I am a chatGPT Plus subscriber. I am becoming increasingly frustrated with the recently released o3 & o4-mini model’s constant use of placeholder sections in the generation of updated file code.

with the o3-mini when I used the following prompt:

Go ahead and generate the complete enhanced code for the file to meet the requirements.

Ensure NO CURRENT functionality is removed.

Do NOT USE PLACEHOLDERS.

Provide FULL COMPLETE error free code as per the requirements.

It would faithfully reproduce the file code with the newly generated code integrated. But now, try as I may it is constantly leaving out major section of file code because it has “not changed”. This is making the models almost unusable!!

I have the same issue, and it seems many other users have the same problem.
It makes o3 and o4-mini-high unusable for coding.

I find that after a very short period, even when reminded of the initial instructions not to do that it will use placeholders instead of full code back-to-back one message after being informed not to do it. I use it to verify my code because I have pretty severe dyslexia and I can’t often see when I get letters in the code switched around. ChatGPT has been a fantastic tool to help me debug issues and to assist in rebuilding my coding skills up until this point. Now it’s more of a hindrance than a benefit.