Card #396 · OpenAI Platform & APIs
Degradation of GPT-5.5 performance and reliability
We are experiencing a significant drop in output quality and reliability with GPT-5.5, making it unusable for serious development. This is accompanied by increased context usage and a lack of transparency from OpenAI, leading some of us to switch providers.
Pain score
0.99/ 1.00 · weighted product of four components
Reach1.00distinct authors (log-normalized)
Recency1.00freshness decay, half-life months
Engagement0.96upvotes + comments, normalized
Monetization1.00willingness-to-pay cues in evidence
§ Evidence — 4 verified quotes
The outputs have become so unreliable that we have had to pull back on development for the time being.
well this explains why 5.5 xhigh has been so horrible for hte past month
With the upcoming deprecation, we are thinking seriously on switching providers
At this point, the business accounts feel like cash incinerators when used for this workflow, because we are discarding most of what is generated.
Comments are a Pro feature — join the conversation, share what you've shipped, hear what others are building.
Upgrade to Pro