Fixes are free, why is review not free?

v0 currently advertises that all errors created by v0 will be fixed for free, which is an extremely important feature for anyone working on complex projects. However, during every generation, v0 enters a stage called “Reviewing Work.” During this stage the agent evaluates its own changes, checks for internal errors and bugs, and then attempts to correct them within the same prompt. While this sounds helpful in theory, the problem is that these internal review and repair operations are counted as part of the paid prompt. This increases the price of the request, often pushes the prompt into a higher cost tier, and in many cases causes the generation to crash entirely. The combination of long review cycles and high compute usage leads to elevated prompt costs and significantly longer work times, even when the user did nothing to trigger these errors.

To maintain fairness and to match what v0 publicly promises, the system should complete the initial user prompt and then begin a separate, free review process that carries over the necessary context. Since v0 states that fixes for errors created by the agent are free, the internal “Reviewing Work” step should not consume user credits or inflate the cost of a prompt. Allowing the agent to review and fix its own mistakes in a separate, cost-free stage would prevent unnecessary price spikes, reduce system crashes, and create a more reliable experience for all users. If this adjustment is not made, the current behavior creates a situation where users are charged for corrections that v0 claims are free, which risks being interpreted as misleading or inconsistent with the advertised feature set.

I am forced to stop reviewed work as it is a big reason v0 stops on longer chats!



I also attached images of the pricing im currently experiencing. These prices are outrages and at a large due to the ‘Reviewing Work’ as reviewing work attempts to redo the entire prompt. Causing either a large spike in cost or a crash, both losing the user credits and money. (Would love a refund on my credits if possible!) they got thrown in the :wastebasket: Thanks v0 team!

1 Like

The team will look into it asap

1 Like

The review step is a bandaid for the fact that we can’t shove unlimited context into a single prompt, so a second prompt is made with additional context to ensure that the generation is still aligned with it.

Almost all of that cost is passed directly to the model provider so it wouldn’t be viable to offer half of it for free, this step is categorically different from the free Fix in v0

But there is definitely room for improvement in large project handling and the team is actively working on them

2 Likes

Totally understandable. I’ve already seen a massive decrease in the “Thinking” cost. Before this update, my prompts were charged as shown in the images above. Today, prompts with the same level of difficulty and containing even more context are being charged about 40% less.


This is amazing considering these were the prices yesterday!

1 Like