Anyone facing antigravity slow. I been noticing that antigravity is lot slower. It take 5 or more than 5 min to complete the simple task. Mostly i am doing myself because is faster. but anyone facing the same issue.
Same issue here, its so slow.
absolutely. it takes 15 minutes for a simple task and it finishes with an error, sometimes. the problem seens to be gemini 3.5 flash. claude is working regularly. did not try gemini 3.1 pro.
update: tried out 3.1 pro, seems to work with normal speed.
Yep, me too. I keep getting a message saying our servers are experiencing high demand. 
Super slow today and buggy
I bet they’re trying to reduce the token usage again until we can no longer use it even on subscription plans lmao
Yep, extremely slow with random 5-10 min hangs on random steps here too, then just when you think you may get at least one task completed the agent gets remotely killed “due to high demand”. This is 18% of my 5 hourly quota down the toilet without seeing a single task to actual completion. How is this acceptable?
If you are compute constrained, at least implement a queue to start new ones and prioritise finishing already running ones!
Really hope they are doing another global reset once things stabilise due to how messy today is as it looks like I won’t get much done as things are and still lost weekly quota over it.
frustrating slow woring write now in balkans eu
I’ve had same problem with Flash (High) today, been using pro all day because it’s running at the same speed.
All of the Gemini models work much better when they are used on platforms other than Antigravity. For instance in OpenCode or Devin (Windsurf) I’ve had very good performance with them. And the Flash models are a good bit cheaper than the mainstream frontier models. Even gemini 3 flash, when used on medium thinking, is quite capable. Cheap and much better than ANY of the Chinese models. No I don’t work for google, I just have been trying out lots of different platforms
ZED is an excellent light IDE alternative and has extension for all the major providers. Much faster and uses far less RAM than VS Code based IDE’s. I’ve been using gemini models through Devin with excellent results. No crashes or service failures. That seems to be an Antigravity problem
Our company is going through the exact same issue. We used to use Antigravity 2.0 with Gemini Pro 3.1 High. This week, we had to migrate to Google’s enterprise plan, which forced us to authenticate using a Google Cloud project.
Later, while testing, we noticed that the Gemini Pro 3.1 High model took up to 3 minutes to run a commit skill—where the job is just to process a git diff and write a conventional commit in English. This used to be super fast, but I don’t know what they’ve done because now it’s incredibly slow, and we no longer see it as a viable tool for our day-to-day work in the office.
I am finding the same
Same problem here, Antigravity 2.0 is very slow.
I was using Antigravity for a long time without issues.
All the problems started after the forced plan changes, which requires me to log in with a Project ID.
Symptoms:
- Tasks got stuck on “Working…”
- Error: RESOURCE_EXHAUSTED (code 429): Resource exhausted.
- Sign-in is lost after a couple of hours working or just after a restart (trying to fix the previous error)
- Changing between the models does not fix the problem, as some people suggested.
- When using the Gemini 3.5 Flash (High), Error 429 happens more often.
- Restricting the key as suggested, did not fix the error or the slowness.
Unusable today, so slow !
I picked gemini 3.5 flash for its speed, but if it’s that slow, no point to continue for LLM less powerfull than Claude !
This is ridiculous - I paid a year subscription. Mid way into it I am now stuck with this unusable cli
Same. Started yesterday and still pretty slow. All infrastructure and services seems ok (no outage).