I found ChatGPT Plus with 5.2 occasionally so stupid it gave me pause, lately more often. I dropped subscription, moved to Claude and was amazed how smart it was. Then realised I’m hitting ceiling after 10 minutes. Back to OpenAI. F*cking hell.
User ditches ChatGPT Plus after Claude 4.6 flawlessly polished a 10k-word legal draft from PDFs.
Deep Dive
A legal professional's viral post details switching from OpenAI's ChatGPT 5.2 to Anthropic's Claude 4.6. While ChatGPT 5.2 made critical errors on a complex 10k-word document, Claude 4.6 (free version) ingested two PDFs and polished the draft to near completion with almost zero mistakes. This real-world test suggests Claude may excel at deep, cohesive analysis of long, complex documents, highlighting a key competitive difference.
Why It Matters
For professionals handling complex documents, the choice of AI model can significantly impact output quality and workflow efficiency.