A tale of LLM competition from a personal usage POV:

I had to cut back on using Claude's Opus model from the API dashboard because of its steep costs.

On the dashboard, Opus had become my go-to, clearly superior to GPT4 and Gemini Pro despite the limited interface...