I'm a heavy power user of language models, especially for tasks like large context coding and long-form story writing. I'm curious if Claude Pro still has the same issues with rate limiting and overall performance. For someone who uses LLMs extensively, is it possible to get a refund on the $20 plan, or would it be more wise to switch to using the API instead? Also, I've heard about alternatives like Poe—what are your thoughts on the best current options for Anthropic Models?
5 Answers
I'm planning to ditch my Claude Pro subscription. Since the recent update, it's been less reliable—you know, ignoring prompts and forgetting context. I sometimes hit token limits with just a few prompts. I definitely miss the performance of earlier versions; it feels like a step back.
The $20 plan can be quite limiting in terms of usage, but it's not all bad when you compare it to the API costs. A single Opus query through the API could set you back around $3-4, so the plan can still be worth it depending on your needs.
I tried the newer Opus model, but I kept running into out-of-memory errors, causing my browser to crash. I've decided to hold off on purchasing any packages for now, as it just doesn't handle the workload well.
If you're focused on coding and app development, you might find Claude Pro lacking. However, I've got a friend who uses it mainly for writing and she swears by the Pro version—it works great for her!
I see both sides. Comparing Claude Pro and Cursor, I find Claude gives better quality outputs and understands my requests better. However, those rate limits are frustrating. The message length limits require me to keep hitting 'continue' and the overall usage cap hits way too early for me. I love the quality, but I'm struggling with these constraints.
Related Questions
xAI Grok Token Calculator
DeepSeek Token Calculator
Google Gemini Token Calculator
Meta LLaMA Token Calculator
OpenAI Token Calculator