May we suggest the pro+ plan?
And coming soon the premium pro+ plan... Then I'm sure we'll see the enhanced premium pro Plus plan...

300 per day is ok, per month is ridiculous

Given Google's Gemini coding agent offers 1500-2000 requests per day on their free tier, I'd say 1000 "premium" (I'd argue - minimum required for agentic coding etc...) requests per month on a paid enterprise plan is most certainly not enough.

Cursor works better, is smoother and costs half as much as copilot. You get 500 requests I think, but extras are only $0.04 each - basically after you work through your $20 of prepaid credits, you can start paying the same rate for any overage.
These companies have no idea how fast this is all happening. Next month we'll have two new options, then 200. We are currently living in the profitable "$.75 per SMS" era of AI.
Prices will either skyrocket out of the hands of most individuals or bottom out and level the playing field.

Either way there are better options than Github Copilot where you don't have to click continue every 10 seconds or try telling the thing to get to work after it's confirmed it's getting to work 6 times.

Considering you're also being billed for requests that had buggy code then you PAY AGAIN AND AGAIN to fix it, and copilot knows this, those additional requests where it says "your right i stuffed up" should be FREE, along with code that is complete garbage or does not follow instructions should be free. (it ai admits this, so then it should tag that request AUTOMATICALLY as 'wrong code' (rather than me trying to do it when i remember (happens so much i've started using thumbs up when it's right.

At one stage i had copilot give me the same 3 replies to the same query three times in a row, and asked it to /fix it in three different ways and it was like a goldfish and gve me the same three answers one at a time three times in a row. That's 10 claude requests for ONE answer. (and code was wrong,so probably 12-16 requests for ONE correct answer.

I'd say 70-80% of requests should be FREE at the moment.

There also needs to be a better triage system for requests so when you ask something simple it asks chatGPT then if it's complex it passes it off to Claude or o4-mini if it's a repetitive simple coding task. That would reduce most people's bills by 90%

Hi @gadfly3173 This could be a feature deployment delay, IDE restart might fix this.

This is my third year using Copilot, but to be honest, while more unnecessary services and features have been introduced, placing a 300-request cap feels quite unfair.

Since it doesn’t seem like this change will be rolled back, I’m considering switching to Cursor once my current annual Copilot subscription ends.
If you have any other recommendations, I’d be happy to hear them.

@hayoshii An alternative for me would be to use local models with Ollama + VSCode or Cursor.

@raky291 Thanks for your input!
Of course, I’ve been experimenting with local LLMs as well.
At first, I only intended to use them as a supplement to GitHub Copilot… but here we are.

For API usage, I'd recommend the anthropic API directly for claude, otherwise openrouter is great for switching between whichever model you want to use.

If you want purely subscription based, not API billing, then claude code with a claude pro or higher sub, or else go local with something like ollama.

@nijaru Thanks! Claude Pro is definitely on my list of options, but OpenRouter sounds especially interesting.

Sad if github copilot was your "wings"...

GPT-4.1 and GPT-4o offer unlimited usage for the Pro Plan, while other models such as OAI, Claude, and Gemini have specific weights for premium request usage. For, using Claude-sonnet-4 (1x) allows 300 uses, while o4-mini/o3-mini (0.33x) permits 900 uses under the Plan, which includes 300 premium requests.
The Pro+, on the other hand, provides 1500 premium requests.

I think there's a misunderstanding somewhere. Gpt4.1 is not unlimited when it comes to premium requests, only chat.

On pro/pro+, you are still limited to 1500 premium requests even with gpt4.1.

I was able to use gpt even when my premium requests ran out with $0 budget (but had to increase that to complete the code ai was helping me with)

also sick of requests that 'timeout' so they can use more credits for long running tasks

Hi @lucvutuan Could you please restart your IDE and check again? The Total Monthly Quota should now be updated with the monthly premium request allowances in effect. Thanks!

Hello @lucvutuan Can you also provide a screenshot with the timestamp ?

And the size (number of tokens) isn't taken into account so tiny requests are billed the same as huge requests - also there's no prompt caching so you're paying full price for all requests.

From my experience, Jetbrains Junie and Claude Code are working great.
I had also tried copilot for VSC and Jetbrains, but it is simply garbage in comparison, not even close.

Others are recommending cursor as a good alternative, but I miss a lot of code features which jetbrains offers in comparison.
You can for example use cursor as ai companion and debug/code with another IDE.

At the moment, I am waiting for a license to test Mistral-Code, to be able to tell which one is best suited for our company.
I hope this helps at least a little. :)

scratch that, cursor also has these stupid premium pricing limits. 500 requests gone in 3 days.

Having a hard time understanding this. 1000 free requests a day is too expensive?

I don't know, within 3h I got a bill for almost 80USD.

Fascinating. I will have to check it out

You get the free request as long as you login (/auth) using your email, not your api key. You can set it up to spend $0 though. Also, this free request ONLY works for the email login I found out.

When I simply logged in, I almost immediately got a notification that the funds were exhausted; so I connected the API and instantly got a notification that the budget was exhausted (I don't know, maybe an hour had passed), so I closed the billing account at the stage of this 80USD.
I am now trying to reproduce this behavior (not using the API key), we'll see if it happens again

[edit]
At this point I have not been able to reproduce this behavior, I still have access in GoogleCLI using login, without API.
Of course, Google charged me money for using the API.

I think they are listening - my bill is charged in a completely acceptable way. i am no longer looking for another solution - at least for now.

How did you get satisfaction? It's 10am on the 30th and I'm supposed to get my 300 requests reset today but I'm still waiting... this experience is amatuer-hour and I feel like the people running this show are clueless corporate drones with no idea what AI even is.

> How did you get satisfaction? It's 10am on the 30th and I'm supposed to get my 300 requests reset today but I'm still waiting... this experience is amatuer-hour and I feel like the people running this show are clueless corporate drones with no idea what AI even is.

Sorry, but I got confused - I was referring to the rate for premium queries in Github Copilot Chat (I have subscriptions for 10USD) and I don't understand 300 queries, as I think it refers to GeminiCLI?

You have 300 requests in github copilot for $10. Every time you touch the interface, talk to the thing, ask a question, answer a question, tell it to keep trying or continue, is a request. See you in a couple days. ;)

Currently, 19% of premium inquiries within the free limit, and I have auto acceptance enabled and practically from 8 to sometimes even 22 I direct the agent's activities.

You know, we can't have our calculators displaying 666, some people might be offended and even be hurt by seeing this satanic number in front of them. So from now on, if the calculator arrives at a result that contains a visible 666, it will simply stop working and refuse to display the number.
We want our calculators to be Responsible™️, but since they're not able to, we'll just hinder parts of their functionality. Sorry if this means somewhat lower quality of your customer experience, but imagine how truly horrible and devastating it would be if someone was just doing their daily math and entered 333 * 2.

Same boat as you, the API are limited to 28 days by design. We implemented our own batch to save all data every day to keep a longer history

This is a different topic (CoPilot metrics through API); I'm talking about premium requests usage report which is available only in UI and only for the current month which is totally suboptimal.

Ah yes these CSV files are a pain in the a**

These people have no idea wtf they're doing other than increasing profits. They're destroying their own profit engine from inside with these shit features and poor analytics. FFS the product ITSELF can do a better job of analytics than the website.

I crashed twice since I posted this, so I guess there goes 10 more!

Yeah I'm crashing constantly now as well. Microsoft is a joke.

It crashes in the middle of edits and then rolls back like 20 requests. It is just wasting my subscription