I've been using Claude for a while now because it helped me with a tricky API integration that ChatGPT couldn't handle. Since then, I haven't really tried other models. I'm curious to know what it would take for you to switch your current AI model. Do you have a specific method or criteria for evaluating them? I still use OpenAI's API for chat completions because it's cost-effective and fast, but when it comes to coding, Claude is my top choice.
5 Answers
I totally get that! I recently tried SWE with Windsurf, and it was just okay. But Claude 4 Sonnet? Pure genius! I sometimes feel like it understands me. That said, I’ve switched to other models when Claude has been slow; Gemini 2.5 has been my reliable backup.
Honestly, I'd need more faith in another lab seriously addressing AI risks. Other companies just seem to market themselves better without genuinely pushing for safety. If Anthropic falls behind, I'm likely to switch to Google, but it would take a lot to convince me to pay for OpenAI or XAi.
I've got some evaluation methods like LMArena, but honestly, I think it’s best to just try them out and see if they suit your needs. For coding, I rely on Claude, but for writing, I mix in Claude and Gemini. And for image reading, I still prefer GPT.
I like to keep tabs on various models out there to see how they stack up. I sometimes switch my subscriptions if I'm not happy with my current model's performance. If the outputs get frustrating, I’ll look at alternatives.
I often switch models on the fly depending on the tasks. It helps to compare outputs and pick the best one each time. I also check out Aider’s leaderboard for a more objective view of which models are performing best.
That's a fair point! I’m not really convinced we should trust companies that are so eager to take military contracts for 'AI safety.' It feels pretty sketchy to me.