Let's address the question directly: Claude (Opus) is not automatically the best choice for Excel formulas and VBA just because it excels at general coding. The user's own experience, finding Gemini superior for this specific task over several months, carries more weight than a reputation that was built on Python, JavaScript, or other programming languages. Spreadsheet logic is a different beast. It operates within a constrained environment of cell references, named ranges, and nested functions that demand a particular kind of structural thinking. A model that shines at building a web app from scratch may struggle to reason through a multi-condition INDEX-MATCH or a recursive VBA loop that needs to handle real-world edge cases.
What this means for you is practical: test before you commit. The user is paying for two subscriptions and still wondering if a third would be better. That is a reasonable frustration, but it also points to a deeper truth, there is no universal champion. Each AI has strengths shaped by its training data and architecture. Gemini appears to handle the spatial reasoning of spreadsheet grids well, which makes sense for formulas that depend on relative positions and range intersections. ChatGPT may offer stronger conversational debugging when you need to describe a problem iteratively. Claude might produce cleaner VBA syntax but could miss the business logic that makes the macro actually useful.
The takeaway is not which AI to pick, but how to evaluate them. Run a controlled test. Take three of your most painful existing formulas, the ones you dread maintaining, and ask each AI to rewrite them. Compare not just whether the output works, but how readable it is, how well it handles errors, and how easily you can adapt it next month. That test will tell you more than any forum post. The user is already doing the right thing by paying attention to real performance over time. Trust that data. Your workflow deserves a tool that earns its place, not one that arrives with a reputation.