Premium plan what to expect slow requests after 500 premium requests?
I think it still a bit unclear from the pricing page.
If i understand correctly it will revert back after 500 premium requests like claude to gpt-4o-mini or cursor-small
Will gpt-4o-mini or cursor-small be of any use at all?
Is it better or the same as gpt4? the mini makes it sound is less good maybe even useless so than the 20 dollars for 500 premium requests is still kinda expensive. I ran a few migration tries to convert nextjs to remix and it just miserable fails with a lot of errors and issues so i ran true the requests really fast.
GPT says:
Here’s an overview comparing these models in plain text for a forum post:
• gpt‑4o‑mini:
This is an experimental, research-oriented model that aims to mimic the outputs of larger systems like GPT‑4 while being more cost‑efficient. It’s estimated to have around 6 billion parameters and typically supports about 2,000 tokens of context. It offers generally coherent responses but is still evolving in terms of performance and reliability.
• cursor‑small:
Designed as a lightweight option, cursor‑small is optimized for speed and lower resource usage. With roughly 2 billion parameters and a shorter context window of about 1,000 tokens, it performs well in fast, less resource-intensive tasks but may struggle with complex reasoning or extended dialogues.
• Claude 3.7:
This flagship model from Anthropic has an undisclosed parameter count, though it’s likely very large. It is known for its strong conversational abilities, advanced alignment, and robust ethical safeguards. Claude 3.7 can handle nuanced instructions and, in some versions, offers extended context windows (up to approximately 100,000 tokens), making it well‑suited for in-depth conversations and detailed tasks.
• Sonnet:
Specialized for creative applications, Sonnet is tuned to excel in creative or poetic generation. While it typically supports a context window of around 2,000 tokens, it may not perform as strongly on technical or heavy reasoning tasks compared to the other models. Its focus is on producing stylistically rich and imaginative outputs.
These comparisons are based on early benchmarks and community feedback, so figures like parameter counts and context lengths are approximate. Each model has its unique strengths: while the mini models focus on cost‑efficiency and speed, flagship models like Claude 3.7 bring robust performance and safety features, and Sonnet caters to creative tasks.
Any real live expierinces with the gpt4 mini on cursor will it be any good? else i just have to cancel my subscription the only thing that copilot cannot do which i have now is write multiple files so its a copy and paste ■■■■ which they would finally solve that. /Im normally using intellJ and really dont like the vscode interface dont know the shortcuts etc.