just experimenting so far but Claude, Grok and Perplexity each one for various projects and use cases have all been very productive. The other LLMs were not on par.