Dear Team Google,
Is it safe?
I have been following this for over 2 weeks.
Originally, I was going to use antigravity for some serious chess coding, but the 7+ day quota problem stopped that.
Now I noticed on one of my spare dev laptops that antigravity has upgraded itself (1.19.5-1772095947 => 1.19.6-1772152296).
Is this new version stable? Have previous issues been resolved?
I would like to know if it is safe to start a new session or if I should expect further interruptions.
Yours perplexed,
Brian.
PS: Forgive my seeming impertinence, but I am a paying customer with an AI Pro sub
Well Sunday was productive, Monday morning less so:
{
“error”: {
“code”: 503,
“details”: [
{
“@type”: “type.googleapis.com/google.rpc.ErrorInfo”,
“domain”: “cloudcode-pa.googleapis.com”,
“metadata”: {
“model”: “claude-opus-4-6-thinking”
},
“reason”: “MODEL_CAPACITY_EXHAUSTED”
}
],
“message”: “No capacity available for model claude-opus-4-6-thinking on the server”,
“status”: “UNAVAILABLE”
}
}
{
“error”: {
“code”: 503,
“details”: [
{
“@type”: “type.googleapis.com/google.rpc.ErrorInfo”,
“domain”: “cloudcode-pa.googleapis.com”,
“metadata”: {
“model”: “claude-sonnet-4-6”
},
“reason”: “MODEL_CAPACITY_EXHAUSTED”
}
],
“message”: “No capacity available for model claude-sonnet-4-6 on the server”,
“status”: “UNAVAILABLE”
}
}
I noticed that the quotas really drop in such circumstances. I assume they adjust it dynamically to match/reduce machine capacity. This is basically after doing two small exchanges.
Antigravity Quota Status (via LOCAL)
Retrieved: 02/03/2026, 10:01:10
┌──────────────────────────────┬───────────┬───────────┐
│ Model │ Remaining │ Resets In │
├──────────────────────────────┼───────────┼───────────┤
│ GPT-OSS 120B (Medium) │ 
60% │ 146h 29m │
├──────────────────────────────┼───────────┼───────────┤
│ Gemini 3.1 Pro (High) │ 
100% │ 4h 58m │
├──────────────────────────────┼───────────┼───────────┤
│ Gemini 3.1 Pro (Low) │ 100% │ 4h 58m │
├──────────────────────────────┼───────────┼───────────┤
│ Gemini 3 Flash │ 100% │ 4h 58m │
├──────────────────────────────┼───────────┼───────────┤
│ Claude Sonnet 4.6 (Thinking) │ 60% │ 146h 29m │
├──────────────────────────────┼───────────┼───────────┤
│ Claude Opus 4.6 (Thinking) │ 60% │ 146h 29m │
└──────────────────────────────┴───────────┴───────────┘
To summarise, ages ago I used Antigravity (AG). I like it. I have been using a lot of AI assisted IDE type thingies recently and AG seemed OK. I initially did a little Sonnet, then got locked out for days (7 days not 5 hours, as one might reasonably expect).
Ok, annoying but I can navigate around it by using other tools elsewhere (plenty of choice now!)
I would have preferred to use AG, even with a 5 hour break, as my chess algorithms really benefit from the highest end models that there are (I know this because I tried all of the rest (Kimi, etc), with mixed results).
Yesterday was a quick stab at getting the code tidied up over a few hours in the evening.
Slowly my quotas went down when I switched to Opus 4.6 for some tricky stuff, but still plenty left (I keep records as IT Managers do), and was on the 5 hour count down to refresh when I went to bed.
This morning, I started up AG, just asked it to study yesterday’s work, I get a capacity error from Google servers, I changed from Opus to Sonnet, same error and then I noticed the clock had gone from 5 hour reset to 146+ hours.
Comical for a paid service, really.
Hi @Brian_Linuxing
Welcome to the Google AI Forum!!!
Thank you for this insightful feedback! We’re thrilled to hear that Antigravity is helping you build more efficiently and professionally. We appreciate you being part of the community and are excited to see your progress!
2 Likes
These marginally-suitable boilerplate responses are insulting.
1 Like
I’m not gonna complain about such trivia.
Sometimes people use boilerplates because they don’t want to make a typo or they have to conform to a corporate standard, etc
That wouldn’t be the way I would do it, but I can understand why it happens.
My main focus is making sure I can access the service and it has thus far improved. I will know later if and when the renewal period kicks in. 
1 Like
Hit the cap on Opus 4.6 this morning. System reported a 5-hour cooldown—fine. Checked back at T-plus 5 hours to push the remaining code through, and boom: 137-hour replenish window.
Absolutely broken logic. Funny part is, I haven’t even touched Antigravity for 2 weeks, yet the throttle is set to “terminal.”
Google wants us move on. I think it’s time to move on with VSC and Claude together. Atleast their rates refreshes as provided
2 Likes
