Has anyone noticed that Gemini 2.5 Pro output code quality has been degraded significantly over the past few days. Model is also ignoring instructions and not providing code in the same format. Due to this drop in quality, our code parsing logic is failing consistently, as the LLM was not adhering to the strict instructions set in our system prompts. Any clue how should we fix this?
1 Like
yes the whole api is messed up these last few days i have a couple of apps that use gemini 2.5 pro and they are all blocked because of this also alot of users reported this
Thanks for your response. Are you aware if Google has acknowledged it and are they working on a fix?
nope no response even on ai studio api status it says it has no problems