Today I experimented with agentic development with Cursor after purchasing a Pro+ subscription costing $60 US, which pays for $60 US of tokens. The overall experience was relatively good, but checking the usage list made me think again whether Pro+ would suffice for any large project work.

Claude 4.5 Sonnet is really good, but burns $ at a rather rapid pace. For easy tasks, the custom AI agent Composer 1 by Cursor is rather sufficient.

Also, having long-winded sessions seems to push costs upwards due to ever increasing context.

Letting a second agent review what the first agent did and a third agent implement the requested fixes reduced costs to $0.50 for those steps. Compared to the roughly $20 I spent on the actual coding and debugging, that's a bargain. It was just about refactoring a small project to use a different library as dependency.

In the end, I didn't need to check the documentations of those libraries. But I needed to check the output of those agents very closely. They tend to drop functionality, if they feel it's difficult to maintain it!

#ai #dev

Reply to this note

Please Login to reply.

Discussion

Maybe check opencode.

Yeah, definitely have to check that out. What's your preferred way of using opencode?

With claude models works nicely! I am using that at the moment.

Probably will try gpt 5.2 with it.

Also here on nostr, you can hear other models people test, you can try those as well.

Opencode is quite popular here.

I hear you. Cursor has one interesting advantage though: If you get the Ultra subscription, you get twice the amount of credits per $, so you go all out on rapid prototyping with premium models.

So, I see myself ending up using a combination of Cursor and Opencode, depending on what I am trying to build during the respective month. After all, paying for Ultra every month feels wasteful.

You can also get a claude subscription to fix your monthly cost, that you use with opencode.

True. The Max 20x plan costs just as much as Cursor Ultra. But the latter allows for all models.

Tough call. Probably depends whether I will be more productive with Cursor or OpenCode. Not sure about that. Have to experiment a lot, first.

Try it. If you can, try multiple combinations.

The agent who writes the code cannot be the reviewer.