Hello - does anyone else have issues with the responses API truncating / merging partial sentences? It doesn’t happen 100% of the time, but I’d say 50% of the time, the output doesn’t make sense. I’ve tried prompting it to “double check the output and make sure it uses complete sentences and doesn’t truncate words” to no avail.
YES ! I have this too : Funny words and missing letters in gpt-5-mini response
I’ve tried other models and other parameters but the issue is still there. I’ve contacted the OpenAi support but no solution so far…