it’s been about two weeks since Gemini started hallucinating cutting code blocks in answers, adding backslashes to code , etc… I’m surprised that i don’t see anybody else complaining about it ?
Problem is i started my career around the release of 2.5, and given the short timeline i anticipated ahead, i completely surrendered to the paperclip maximizer and i’m vibe coding at work. Gemini down = man down. It’s a serious matter.
Hello @bk55454, to help us investigate this issue, could you please share the exact prompt and any system instructions you are using so we can attempt to reproduce this on our end?
Flash-Lite seems to apply a correction (the first output is wrong and is then sanitized).
Couldn’t reproduce the bug with models other than 3.6 reasoning on the Pro/enterprise version.
