I can confirm that. Especially since yesterday (26.03.) evening Gemini Code Assist slowed down dramatically, even on new chats and now refuses to work at all. First info in the chat is the fallback to 2.5 (which is not really suitable for the size of my project) and then not even 2.5 creates any output anymore.
I think Google has a serious issue here and it would be great to get some information/updates when this will be fixed.
Hi @pallabmukherjee3
For what model are you experiencing this ?
@Mustan_lokhand
I have the exact same request
I was using the gemini CLI today on my AI Pro plan and got throttled so hard that one request got throttled for 30 minutes
I got this snippet from the debug console there:
' "status": "RESOURCE_EXHAUSTED",\n' + │
│ ' "details": [\n' + │
│ ' {\n' + │
│ ' "@type": "type.googleapis.com/google.rpc.ErrorInfo",\n' + │
│ ' "reason": "MODEL_CAPACITY_EXHAUSTED",\n' + │
│ ' "domain": "cloudcode-pa.googleapis.com",\n' + │
│ ' "metadata": {\n' + │
│ ' "model": "gemini-3.1-pro-preview"\n' + │
│ ' }\n' + │
│ ' }\n' + │
│ ' ]\n' + │
│ ' }\n' + │
│ '}\n' + │
│ ']', │
│ headers: { │
│ 'alt-svc': 'h3=":443"; ma=2592000,h3-29=":443"; ma=2592000', │
│ 'content-length': '630', │
│ 'content-type': 'application/json; charset=UTF-8', │
│ date: 'Thu, 09 Apr 2026 18:50:54 GMT', │
│ server: 'ESF', │
│ 'server-timing': 'gfet4t7; dur=7056', │
│ vary: 'Origin, X-Origin, Referer', │
│ 'x-cloudaicompanion-trace-id': '65bd1c35535158af', │
│ 'x-content-type-options': 'nosniff', │
│ 'x-frame-options': 'SAMEORIGIN', │
│ 'x-xss-protection': '0' │
│ }, │
│ status: 429, │
│ statusText: 'Too Many Requests', │
│ request: { │
│ responseURL: 'https://cloudcode-pa.googleapis.com/v1internal:streamGenerateContent?alt=sse' │
│ } │
│ }, │
│ error: undefined, │
│ status: 429, │
│ Symbol(gaxios-gaxios-error): '6.7.1' █│
│ } █│
╰─────────
I started at 0% usage on my quota, and got 20 API request errors with an error rate of (76.9%)
Only 195,536 tokens
most being cache reads 112,400 (65.9%)
Due to retries
I am getting throttled so hard that my plan is unusable. I love the AI pro plan stack and features, but it is coming to the point where it is ridiculous to try and use.
My next step is to not use Google AI Pro, and up my Claude plan, even though I love the Gemini features.
Please help me with this issue
I am facing this issue with all pro and preview models.
Same problem here. For like last 1-2 weeks its a nightmare to use on gemini-cli - 3.1 pro and 3 flash responds usually in seconds, but now it takes like 2-11 minutes… I am out if it is not fixed till next billing period in 3 weeks. Competition doesn’t sleep.
“reason”: “MODEL_CAPACITY_EXHAUSTED” - and i barely opened gemini-cli, with almost 100% limits left. good job google.
I have Gemini Pro subscription. Gemini CLI and Antigravity are taking too long to respond.
I am seeing this issue as well. Even if I change the model to Fast instead of Pro or Thinking , then also the response is generated using thinking or 3.1 pro which is taking too much time. I dont want a pro level analysis for every single prompt. pfa attached screenshot.
I’m unsure if anyone has said this yet, but could be down to the servers. Maybe the time you started to use gemini, many users where using it at the same time?
I’ve personally have had no issues so thats why i have mentioned the above
Hmm, while I haven’t heard anything, I’ve seen a lot of patches though. And, from my point of view, the issues has been solved.
