AI Assistant consuming credits much faster since WebStorm 2025.2.1

Hello David Bennett

It seems we might be experiencing the same issue that was reported in YouTrack. Could you please take a look and confirm if it’s the same one?
https://youtrack.jetbrains.com/issue/LLM-20825

I’ll also contact you directly in the ticket for further investigation.

Best regards, Elina
 

Tokens run out very quickly, and this varies depending on the model. Even if it can't find a solution or throws an error, it still spends significant amounts. I'm restricting the index to prevent it from spending so much, and this time, it gives an error. Jetbrains MCP needs to convert requests into a more appropriate format and forward them to the model to reduce token consumption. If this continues, Jetbrains AI won't be very useful to us.

About 14 hours of coding with Junie, all of my monthly quota on the Ultimate plan has been consumed.

This is ridiculous, I used to be able to go an entire month without ever hitting the quota limit. Getting really tired of Jetbrains trying to find more ways to extract money from their customers' wallets every month. We already pay a subscription fee for the IDE's, another for AI, and now you want us to buy top-up credits just to have the same amount of usage we had all along from the beginning? This feels very greedy and underhanded IMO. 

I've been subscribed to the Jetbrains All Products Pack for almost 7 years… Jetbrains AI since it was first offered. If these are the games they are going to play, I think it's time to look at alternatives like Cursor or Windsurf… 

I totally agree with everyone here-the AI Assistant's credit consumption is crazy fast since the last update. I'm on Ultimate, and what used to last me a whole month is gone in just a couple of days of normal work. JetBrains says they changed the quotas, but honestly, it feels like they made an expensive tool completely unusable unless you want to pay for constant top-ups.

This new burn rate is a huge letdown and makes me regret getting the higher-tier subscription. If this isn't fixed soon, I'm definitely switching to another, AI solution because this is just not worth the money anymore.

Hello, and thank you for sharing your feedback.
We understand your frustration with the faster quota consumption. The change you’re noticing is due to the AI quota model update introduced on August 25, 2025, which was necessary to keep the service sustainable, as we pay LLM providers for usage.

Quota usage can also vary greatly depending on the prompts you use. To optimize results and reduce costs, we recommend first building and testing your prompt templates with a lighter model (e.g., Gemini Flash) — once you refine the structure and flow, you can apply the same prompt to a more advanced model to achieve higher-quality results with fewer tokens.

You can find more details on what may cause high quota usage here:
https://youtrack.jetbrains.com/articles/SUPPORT-A-2448

You can also use local models (which don’t consume cloud quota) or purchase Top-up credits, which roll over month to month. We’re actively reviewing quota efficiency and transparency to make the experience more balanced and flexible for all users.
If you need further help, please don’t hesitate to reach out to our support team: https://jb.gg/ai-assistant-support

you were lucky to get 2 hours, my entire monthly quota on the Ultimate plan went in about 3 chats with very few questions asked.

The quota since August 2025 is a total joke now. 

btw, I'm on CLion/RustRover, not webstorm, but the issue is the same.

As a business owner, I understand making it sustainable - but you do that new subscriptions and renewals on existing subscriptions. Your leadership made two very bad decisions:

1) You didn't think through the implications of the pricing for your offering - you should have realized it was not sustainable because you would be incentivizing lazy vibe coding on the most expensive models. I suspect you just charged an amount for the subscription without tying it to cost of goods sold - another bad decision. Was your CFO not consulted on this? They would have raised these concerns (likely silenced).

2) You decided it would be best to pull the rug from under your customers - not by changing the price at renewal time (that is a time where another sale has to be made by the subscriber deciding if the product is still worth the price for another year), but by retroactively reneging on your agreement. That was a crap business decision, and one leadership wouldn't have made if they weren't blinded by panic in the market.

Lastly - if I have a quota, why are you stealing my quota that I paid for? At least don't put salt in the wound! You would have softened a lot of the criticism if those of us that are intermittent with our use would at least be able to use credits when we have time to use them. My wife just had baby #6 followed by me having surgery and was off work and side projects for 2 consecutive months - I lost 2 months of credits because of your lousy policy.

Used up a third of my monthly credits on the Ultimate tier within 2 hours with zero Junie use. This is hilarious. I've been a Jetbrains customer for years, but I will not be sticking around. Glad to see I'm not the only one.

I created a support thread of my own about AI Assistant consuming credits even the you're not using it at all. I put my laptop to sleep(no background services running or anything) at the end of the day. Next morning I look at it and a few credits were consumed. I doubt Jetbrains will issue credit refunds when they figure out that their consumption tracking logic is completely broken.

I've been taking periodic screenshots of the daily consumption since my quota reset this month. I made sure to disable AI assistant on my projects and I haven't made a single prompt. Somehow almost 2 credits have been consumed since Nov 3, with AI assistant disabled. I'll continue keeping a record this month because credits are seemingly disappearing on their own

Using the Ultimate tier, I was previously unable to burn through ½ my quota. I just noticed, I have 1.9% quota remaining for the month, which will reset in 4 days. This appears to be a HUGE quota change.  I also feel like a “bait and switch” victim. I will be checking the market for other AI options.

Bump.  In recent weeks, I've noticed my credits burning at a much faster rate than before using the same agentic model.  These are basic tasks, no Junie.  I'm also looking at alternatives.

Following up on my previous post - After evaluating a number of other AI solutions, I have switched over to OpenAI Codex…Even on the $20/mo Plus plan, I have yet to hit the weekly rate limit doing more or less the same amount of work I did previously with Junie/AI Assistant. The only caveat with Codex is there is no tab autocomplete - which I rarely used anyway. 

I hope others posting here do not tolerate this shady behaviour from Jetbrains and also look at alternatives.

After 1–2 hours with only a few requests, a quarter of the credits are already used up. Unfortunately, the new model is no longer usable. One should really cancel immediately and ask for a refund. If this doesn’t improve again, the subscription will be cancelled at the next renewal date.

Logged back in to write that I updated the AI Assistant, and my usage is much more reasonable now.  May have been a temporary bug that caused the credit to exhaust so quickly?

I have also seen the AI assistant send the same file about 30 times in one simple prompt.

I think the pricing and subscription options need to change. I converted Kiro to Jetbrains as is :D

I bought 1000 credits for $20 and closed the deal. It's endless, and I highly recommend it.

Hello,

Thank you for your feedback. We sincerely apologize for the inconvenience you’ve experienced with quota consumption.

In many cases, high quota usage is expected — for example, when using more expensive models or when working with Junie, which requires additional resources. However, if this doesn’t apply to your situation and you still notice unusually high quota usage, it may be caused by bugs (for example, LLM-20825).

In such cases, please contact our support team. We will investigate the issue and help identify the cause.

We are also planning to release the BYOK (Bring Your Own Key) feature later this year, which is expected to enhance the quota experience.

Additionally, we encourage you to vote for the feature request to implement a quota usage report - LLM-19959, which will make quota consumption more transparent.

I also upgrade to ultimate package recently, and went from “85% usage” to “no usage for 18 more days” in a single session.

The quota given is absolute trash, I am going to have to go back to copy-pasting directly to gpt in the browser.

On top of this, lately, the groovy support has gotten so bad, I can't even look up references to most things without waiting 10-20 seconds.  I've been a happy, paying customer of IntelliJ ultimate for many years, but this weekend, I am actually downloading eclipse after many years to see if it is a usable alternative.

JetBrains, you guys are going to lose a lot of customers over this.  I don't mind paying a few hundred a year for a better service, but I refuse to pay thousands of dollars for a product that has been getting _worse_ for my use cases over time and is now using exploitative pricing models to bait-and-switch customers.

I disabled AI assistant across all my projects for the entirety of the month of November and all of my credits were still consumed. Jetbrains is essentially robbing us. I took a few screenshots throughout the month as proof.

My credits renewed on December 1, 2025, today is December 8 and I'm already down to 6.22/10.00… again, with AI Assistant still fully disabled.

November 4, 2025

November 5, 2025

November 12, 2025

December 1, 2025

December 8, 2025

Hello cbelt,
Could you please submit an internal support ticket to our support team using the following link: https://jb.gg/ai-assistant-support?
Thank you!

cbelt Shocking!  Can you please let us know the result of your internal support ticket.  I see the same thing with AI features disabled, it slowly ticks away by itself…

@io games It would really help if the IDE showed: per-request or per-session estimated credit cost, and a breakdown of what consumed quota (model, context size, Junie vs regular).

This is basically pushing me to go with Antigravity through the Google One subscription. It’s 20€, and the quota resets every 5 hours or so, a much better deal overall.

I really like WebStorm, but Antigravity has the stronger offer right now.