# Grok Bot weekly usage meter flips between 77% and 100% in real time

**URL:** <https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909>\
**Category:** Bug Reports\
**Tags:** grok-bot\
**Created:** [November 1, 2026, 4:25am UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909 "2026-11-01T04:25:32Z")\
**Posts on this page:** 20\
**Page:** 7

<div class="post-metadata">

**Author:** ![maKSim85](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/maksim85/32/125156_2.png) [@maKSim85](https://forum.cursor.com/u/maKSim85)\
**Post date:** [November 2, 2026, 11:19am UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/159 "2026-11-02T11:19:02Z")

</div>

How the hell can this even be legal? From the EU’s perspective, this shouldn’t be allowed … I bought the $60 plan a month ago just because of the $20 usage fee for GrokBot, and I thought $60 would be perfect for my needs … I paid for a year, and now I’m only getting half the usage… This is a total rip-off …

---

<div class="post-metadata">

**Author:** ![Scott\_Myers](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/scott_myers/32/125071_2.png) [@Scott\_Myers](https://forum.cursor.com/u/Scott_Myers)\
**Post date:** [November 2, 2026, 11:21am UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/161 "2026-11-02T11:21:23Z")

</div>

It’s 100% not legal. He doesn’t give a ■■■■ because any lawsuit that comes he just pays it. There needs to be actual consequences for this ■■■■■ not a slap on the hand. This is ridiculous. I’ll never buy a product from Tesla, Cursor, SpaceX or whatever other screwed up company he works with every again. The bridge has been burned. He’s a con artist and this is just the latest example of it.

---

<div class="post-metadata">

**Author:** ![maKSim85](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/maksim85/32/125156_2.png) [@maKSim85](https://forum.cursor.com/u/maKSim85)\
**Post date:** [November 2, 2026, 11:27am UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/163 "2026-11-02T11:27:56Z")

</div>

bro this was the same with VS Code and Github Co Pilot plan so every ■■■■■■■ US based company does this.

---

<div class="post-metadata">

**Author:** ![Scott\_Myers](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/scott_myers/32/125071_2.png) [@Scott\_Myers](https://forum.cursor.com/u/Scott_Myers)\
**Post date:** [November 2, 2026, 11:28am UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/165 "2026-11-02T11:28:51Z")

</div>

Haha funny you mention that, I cancelled that too because of that exact issue. It’s all a joke.

---

<div class="post-metadata">

**Author:** ![maKSim85](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/maksim85/32/125156_2.png) [@maKSim85](https://forum.cursor.com/u/maKSim85)\
**Post date:** [November 2, 2026, 11:33am UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/167 "2026-11-02T11:33:56Z")

</div>

This post was flagged by the community and is temporarily hidden.

---

<div class="post-metadata">

**Author:** ![DannyB](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/dannyb/32/119004_2.png) [@DannyB](https://forum.cursor.com/u/DannyB)\
**Post date:** [November 2, 2026, 12:52pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/168 "2026-11-02T12:52:04Z")

</div>

Here’s a list of alternatives that I’m thinking of trying. Let me know your thoughts on them 🙂

> [@Introducing The X Subscription](https://forum.cursor.com/t/introducing-the-x-subscription/174450/426):
>
> Has anyone tried replacing Cursor/Grok Bot with Hermes and one of the open models? The usage problem on Grok Bot made me research it and I’m considering installing a model on my local machine. I asked Gemini and it recommends: DeepSeek 4 Meta Llama/ollama AliBaba Qwen/Qwen-Coder Zhipu/Moonshot GLM/Kimi Google Gemma 4 Any I’ve missed to consider? I haven’t tried any of them yet. I’ve also never used Claude Code or ChatGPT Codex. I’ve only ever used Cursor and Grok Bot. Thoughts on how Herme…

---

<div class="post-metadata">

**Author:** ![sxz](https://avatars.discourse-cdn.com/v4/letter/s/a587f6/32.png) [@sxz](https://forum.cursor.com/u/sxz)\
**Post date:** [November 2, 2026, 12:57pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/171 "2026-11-02T12:57:05Z")

</div>

Update Mon Nov 2: my meter now reads 64.2% with no reset in between, which matches my estimate above. But on-demand still shows $8.66 billed while the false 100% was showing. So the usage was corrected, but it wasn’t just a visual artifact.

---

<div class="post-metadata">

**Author:** ![hyk](https://avatars.discourse-cdn.com/v4/letter/h/b487fb/32.png) [@hyk](https://forum.cursor.com/u/hyk)\
**Post date:** [November 2, 2026, 1:31pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/172 "2026-11-02T13:31:24Z")

</div>

Data from my own usage export (Pro+). All times PST.

**Before Nov 1, the meter was consistent:** about 7.3M total tokens per 1% of the weekly limit.

- Oct 18–25: 731M tokens = 100%.
- Oct 25 until the Nov 1 jump: 579M tokens = 79% (then it jumped to 100% with no activity).
- The two weeks match within 1%, on raw tokens and on any reasonable weighting (cache reads cheaper, output pricier).

**Since the reset (6:14 AM Nov 1):**

- Overnight it went up to 14%, then dropped back to about 8%.
- 12:05 AM → 3:32 AM Nov 2: 8.5% → 16.3% (7.8 points) for about 30.6M tokens, or about 3.9M per 1%.
- 3:32 AM → 4:43 AM: another 0.5 points for under 2M tokens.
- That’s roughly 0.45–0.53× the old rate: **about 2× the cost per token, effectively a halved weekly limit.**
- Whole week so far: about 96M tokens = 16.8%, or 5.7M per 1% (about 78% of the old rate).

**Same usage mix.** About 92% of this week’s tokens are cache reads (90% last week), and 92% are the same default Grok Bot model. A cache-heavier mix should make the meter cheaper, not more expensive.

---

<div class="post-metadata">

**Author:** ![Francky](https://avatars.discourse-cdn.com/v4/letter/f/a8b319/32.png) [@Francky](https://forum.cursor.com/u/Francky)\
**Post date:** [November 2, 2026, 2:52pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/174 "2026-11-02T14:52:13Z")

</div>

Open Bot is an interesting one to look at…I just like that Grok Bot is all hosted in the Cloud…Open Bot gives that Grok bot type Ui…but its all run on your Machine locally, which im not sure about yet.

---

<div class="post-metadata">

**Author:** ![Scott\_Myers](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/scott_myers/32/125071_2.png) [@Scott\_Myers](https://forum.cursor.com/u/Scott_Myers)\
**Post date:** [November 2, 2026, 3:06pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/175 "2026-11-02T15:06:25Z")

</div>

You need a GPU, one that can fit an LLM that can use tooling for it to be actually useful. I have LM Studio now with the Locally app on iOS. It works but I haven’t got any tooling working. For example, I need it to update my wiki.js pages, ssh and fix Linux related issues etc. Grokbot excelled at this.

I’m going to look more tonight. I came across open bot as well but I don’t think it has the same functionality.

---

<div class="post-metadata">

**Author:** ![DannyB](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/dannyb/32/119004_2.png) [@DannyB](https://forum.cursor.com/u/DannyB)\
**Post date:** [November 2, 2026, 3:09pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/176 "2026-11-02T15:09:16Z")

</div>

Not sure how good this is @Scott_Myers:

[![](https://us1.discourse-cdn.com/cursor1/original/3X/3/e/3e793b33315091f30114645be90c192e42853839.jpeg "Run 400B+ AI Models on a 4GB Laptop (AirLLM)") ](https://www.youtube.com/watch?v=tqpar1iyxiU)

It claims to be able to run a 400B model on 4GB of VRAM.

I just saw the video yesterday but I’m a noob myself.

---

<div class="post-metadata">

**Author:** ![Francky](https://avatars.discourse-cdn.com/v4/letter/f/a8b319/32.png) [@Francky](https://forum.cursor.com/u/Francky)\
**Post date:** [November 2, 2026, 3:18pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/179 "2026-11-02T15:18:31Z")

</div>

You dont need a GPU…Open bot you can select for example a Claude Subscription or even a Grok Sub and it will use your LLM Subscription to run…but you can also point agents towards a locally run LLM such as Qwen etc, but then you’d need a decent Graphics card…which I have (RTX 5090) but dont want to use it for that…but all the files etc run and are stored on your PC.

---

<div class="post-metadata">

**Author:** ![Scott\_Myers](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/scott_myers/32/125071_2.png) [@Scott\_Myers](https://forum.cursor.com/u/Scott_Myers)\
**Post date:** [November 2, 2026, 3:28pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/180 "2026-11-02T15:28:42Z")

</div>

That’s the problem. I’m sick of paying these companies that will pull their tokens or increase pricing. I can self host entirely with different models. The model is not the issue but the tooling is. For example I had grokbot connecting to my local machines, solving Linux issues, doing backups, basically I was hands off while it did stuff. If open it can do the same the n it might be an option. At work so I’ll look into it tonight. Thanks for pointing it out.

Also Danny, that’s a no go. 3+ minutes for a response is no good and will be even worse with other requests.

---

<div class="post-metadata">

**Author:** ![DannyB](https://sea3.discourse-cdn.com/cursor1/user_avatar/forum.cursor.com/dannyb/32/119004_2.png) [@DannyB](https://forum.cursor.com/u/DannyB)\
**Post date:** [November 2, 2026, 3:31pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/181 "2026-11-02T15:31:02Z")

</div>

I’m keen to hear your findings please @Scott_Myers. I haven’t got the time right now to try it myself.

---

<div class="post-metadata">

**Author:** ![blaze](https://avatars.discourse-cdn.com/v4/letter/b/e274bd/32.png) [@blaze](https://forum.cursor.com/u/blaze)\
**Post date:** [November 2, 2026, 3:39pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/182 "2026-11-02T15:39:13Z")

</div>

Lol so I am the one misleading… @deanrie

 ![CleanShot 2026-11-02 at 10.38.06 AM@2x](https://us1.discourse-cdn.com/cursor1/original/3X/1/c/1c307d1249baefa069f9dcc1d624de08a8ba706b.jpeg)

---

<div class="post-metadata">

**Author:** ![blaze](https://avatars.discourse-cdn.com/v4/letter/b/e274bd/32.png) [@blaze](https://forum.cursor.com/u/blaze)\
**Post date:** [November 2, 2026, 3:41pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/184 "2026-11-02T15:41:42Z")

</div>

Totally agree, we need to be looking to get compensation from Cursor. I email them today and I expect them to issue a prorated compensation on my account; otherwise, I will be switching the moment there is a viable alternative from competitors

---

<div class="post-metadata">

**Author:** ![blaze](https://avatars.discourse-cdn.com/v4/letter/b/e274bd/32.png) [@blaze](https://forum.cursor.com/u/blaze)\
**Post date:** [November 2, 2026, 3:44pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/186 "2026-11-02T15:44:51Z")

</div>

What are they flagging here, @maKSim85?

---

<div class="post-metadata">

**Author:** ![Heroeskitchen](https://avatars.discourse-cdn.com/v4/letter/h/4bbf92/32.png) [@Heroeskitchen](https://forum.cursor.com/u/Heroeskitchen)\
**Post date:** [November 2, 2026, 3:46pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/187 "2026-11-02T15:46:21Z")

</div>

This happened to me over the weekend, I was at 64% then suddenly at 100% until just this morning. Apparently there was a billing and usage support incident that was resolved yesterday:

> **[✓ Service Degradation - Billing and Usage Dashboard - Cursor Incident |...](https://statusstack.com/service/cursor/incidents/01a1959f-b8cb-76f7-9004-b5a3750737ba)**
>
> Resolved incident: Service Degradation - Billing and Usage Dashboard. View timeline and updates for this Cursor incident.

---

<div class="post-metadata">

**Author:** ![CallYourPal](https://avatars.discourse-cdn.com/v4/letter/c/67e7ee/32.png) [@CallYourPal](https://forum.cursor.com/u/CallYourPal)\
**Post date:** [November 2, 2026, 4:18pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/188 "2026-11-02T16:18:07Z")

</div>

**Update:** As of this morning (Mon Nov 2), my weekly usage was corrected from 75% to about 44%, and the meter has stayed steady. That’s almost exactly what I expected: 36.5% Saturday night plus the 7 or 8 points from my actual weekend activity.

Checking against what I asked for:

1. **Metering bug confirmed:** It looks like the team acknowledged it.
2. **Usage corrected:** Yes, back to my real number.
3. **Rate change on older plans:** Still unanswered. A short statement that rates for Pro, Pro+ and Ultra did not change would close this out for a lot of us.

Thanks to the team for fixing it. If you were affected and your meter still hasn’t been corrected today, post your plan, your current reading and your reset time so they can see who’s still waiting. @Brock_Metz, has yours been restored ahead of your Thursday reset?

---

<div class="post-metadata">

**Author:** ![Gak\_Chen](https://avatars.discourse-cdn.com/v4/letter/g/b77776/32.png) [@Gak\_Chen](https://forum.cursor.com/u/Gak_Chen)\
**Post date:** [November 2, 2026, 4:55pm UTC](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909/189 "2026-11-02T16:55:34Z")

</div>

Adding data from my own usage export (legacy Pro+, Grok Bot only, Taipei time):

- Week of 10/23: ~6.9M tokens per 1% of the weekly limit.
- After the 10/30 reset: ~3.7M tokens per 1% (~47% less).
- Today the meter was refunded about 20 points (43.6% → 23.7% between 09:28 and 16:47), so the weekly total now looks normal (~190M tokens at 27.9%, ~6.8M per 1%).
- But usage _after_ the refund is still at the new rate: 16:47 → 21:16, ~14.5M tokens moved the meter +4.2% (~3.5M per 1%); 20:06 → 21:16, ~10.5M tokens for +2.7% (~3.9M per 1%).

This matches hyk’s numbers in #172. So it looks like a one-time refund of the overcharge, not a rollback: we’re not using more, the pool got smaller. Could staff please confirm whether the per-1% rate for legacy plans has changed, and what it will be after the next reset?

[Previous page](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909.md?page=6)

[Next page](https://forum.cursor.com/t/grok-bot-weekly-usage-meter-flips-between-77-and-100-in-real-time/176909.md?page=8)
