I’m comparing OpenAI and Claude for work that involves long reports, meeting notes, and follow-up questions. Both can be useful, but I care more about whether the model keeps details straight across a long conversation than about getting a polished first draft.

Has anyone used both regularly for this kind of work? I’d be interested in how they handle document limits, corrections, and answers that need to stick closely to the source rather than fill in gaps.