Feature-request: Use non-agentic/non gpt-5 version of v0

Previously I was able to create a whole project with 3-5 versions or multiple projects with a 1-2 versions with just $1 of tokens. Now I run through the tokens like crazy. A project costs me $5-10 dollars, basically 10x the cost while I there’s only slight improvements in code (fewer errors and better adherence).

The price/quality might be worth it for someone who doesn’t know how to code, because I admit that the code is slightly better. But for me and many people I spoke to the change to agentic mode skyrocketed costs without significant benefits.

Please allow professionals to use non-agentic versions of v0 that keep costs low, speed high and quality acceptable.

This type of feedback has been listed all over the forums, and literally zero useful feedback on the regressions from any staff or management or what might be improved.

It’s posted all over the forum, I opened a thread as well but it seems like they don’t have any plans to bring back the previous models. Hopefully enough people will submit their frustrations and we could get back the wonderful models that we had before. Agent simply doesn’t work, the previous models were better in every user parameter. I’m sure that they are doing some evals at scale, but if the user experience doesn’t match the evals they don’t matter. I’ve suggested to at least bring back the choice (like Curser) between their previous models (Small, Medium and Large) and the agent.

Yeah I registered a forum account just because of that. I’m sure the staff is reading this and hopefully at some point there’ll be a change. For me personally I won’t renew my subscription and will just use Cursor. If you feed it some screenshots and guide it with the right prompts (what technology to use and how to implement it) I get the same results for the same or less token price.

Vercel Staff

Appreciate you sharing this feature request. This is not a feature I would expect to see added immediately. However, your feature request has been shared with the v0 team for consideration (both this post and previous related conversations).

Feel free to continue the conversation here, but remember to follow our Code of Conduct

reply to #6

Same here, registered only for this feedback. Seems like it’s at least at the stage that it’s being considered as a feature request. I think it’s important to point out that we want the previous models: Small, Medium and Large - exactly how they worked. Not new models, or a different type of model but what was previously available it was years a head of everything that is currently exists in the market.

reply to #4

Apparently my feedback is being hidden now too - so criticism violates their guidelines too...

Hello All,

Any updates here? or possible timeline? Working with V0 became pretty rough, to the point where we are currently using Figma.

Any timeframe here please? Here is an example of a chat today id qZ9j7JNaaPW A basic request is performed really bad, anything that requires any nuance takes way too many prompts.

Vercel Staff

The feature request has been shared with the team, but no immediate plans to implement given other roadmap items in progress, so I'm afraid I don't have a timeframe to offer

This topic is closed. It no longer accepts new replies.