I'm really fed up with Claude and its constant problems. The Claude models have disappointed me for a while now, especially with their latest updates. I used to like Claude 3.7, but with the switch to Claude 4 Sonnet and Opus, the quality has noticeably dropped. I thought the context limit was supposed to be 200,000 tokens (around 150,000 words), but I've never come anywhere close to that. I've hit their maximum length well before even finishing my stories even when they are only about 50,000 words long.
Another major issue is that I often face daily limits when trying to correct mistakes made by the AI, which can be really frustrating. I'm on the Pro plan and the errors it makes mean I just run out of messages trying to get decent outputs, which feels unfair. The latest update now even has Claude mixing languages in frustrating ways, which has never been a problem for me before.
I'm seriously considering ending my subscription. Even though Claude has been the best model I've found so far for creative writing, there are so many issues that it's becoming unbearable. I'm looking for recommendations for alternative AI that can provide similar or better creative writing without all these drawbacks. I know Gemini 2.5 has some decent features, but it has its own issues too, especially regarding chapter length. Does anyone have suggestions for alternatives that can deliver quality creative writing without these hassles?
1 Answer
Gemini Pro 2.5 tends to do a solid job with critical feedback and rewriting, especially for children's stories. Plus, I've been using OpenAI's GPT-4.5 lately, which is said to be good for emotional depth. I usually run a story through Claude and then send it to GPT-4.5 for a rewrite—might be worth a shot!
While I get that GPT-4.5 has its strengths, I’ve found it pretty limited too—censored and often produces really short chapters (around 500 words max). Gemini Pro seems better in comparison, though it’s not without flaws. Have you tried any other LLMs?