Update break model

Google Antigravity
,
1

After updating, none of the models work and it keeps returning this error in a loop.

image
image440Ă—247 7.94 KB

2026-03-07 07:49:42.800 [info] I0307 07:49:42.800197 34576 planner_generator.go:320] Encountered retryable api error. retrying in 5s. Error (The model is currently unreachable.): UNAVAILABLE (code 503): No capacity available for model gemini-3.1-pro-high on the server: UNAVAILABLE (code 503): No capacity available for model gemini-3.1-pro-high on the server

2026-03-07 07:49:47.800 [info] E0307 07:49:47.800247 34576 log.go:398] UNAVAILABLE (code 503): No capacity available for model gemini-3.1-pro-high on the server: UNAVAILABLE (code 503): No capacity available for model gemini-3.1-pro-high on the server

2026-03-07 07:49:48.053 [info] I0307 07:49:48.053280 34576 planner_generator.go:285] Requesting planner with 9 chat messages at model retry attempt 1 and API retry attempt 2

2026-03-07 07:49:53.564 [info] I0307 07:49:53.563856 34576 http_helpers.go:123] URL: https://daily-cloudcode-pa.googleapis.com/v1internal:streamGenerateContent?alt=sse Trace: 0xd3460d1d1a7e0629

2026-03-07 07:49:53.863 [info] E0307 07:49:53.863130 34576 log.go:398] agent executor error: could not convert a single message before hitting truncation

2026-03-07 07:49:57.870 [info] E0307 07:49:57.870679 34576 log.go:398] agent executor error: could not convert a single message before hitting truncation

2026-03-07 07:49:58.780 [info] E0307 07:49:58.779895 34576 log.go:398] agent executor error: could not convert a single message before hitting truncation

2026-03-07 07:49:59.228 [info] E0307 07:49:59.227809 34576 log.go:398] agent executor error: could not convert a single message before hitting truncation

2026-03-07 07:49:59.405 [info] E0307 07:49:59.405738 34576 log.go:398] agent executor error: could not convert a single message before hitting truncation

image
image924Ă—602 25.7 KB

1 Like
2

happen to me as well, currently paid for Ultra so sad have work to deliver :frowning: . this need to be fix ASAP

1 Like
3

Updated today and keep getting Error Unknown: Agent execution terminated due to error. on loop for all models. Then when you look into the logs it is the above. No matter the size of the prompt or model, this happens. No options provided to clear the cache or memory. No options for fixing this error. I am going to have to switch my projects over to another AI builder because I need to be able to work unless I can fix this soon.

Antigravity Version: 1.20.4

4

Try downgrading to 1.19.6.

5

Today’s update 1.20.5 still doesn’t fix the problem.

6

use this command in powershell for downgrade easy

winget install Google.Antigravity --version 1.19.6 --force

4 Likes
7

Trajectory ID: 35526a30-f0ee-4952-8a6f-2a2ce01c1…
Error: agent executor error: could not convert a single message before hitting truncation
(1) attached stack trace
– stack trace:
| google3/third_party/jetski/cortex/cortex.(*CascadeManager).executeHelper.func1
| third_party/jetski/cortex/cascade_manager.go:2096
| […repeated from below…]
Wraps: (2) agent executor error
Wraps: (3) attached stack trace
– stack trace:
| google3/third_party/jetski/cortex/chatconverters/chatconverters.init
| third_party/jetski/cortex/chatconverters/trajectory_chat_converter.go:44
| runtime.doInit1
| third_party/go/gc/src/runtime/proc.go:8105
| runtime.doInit
| third_party/go/gc/src/runtime/proc.go:8072
| runtime.main
| third_party/go/gc/src/runtime/proc.go:258
| runtime.goexit
| third_party/go/gc/src/runtime/asm_amd64.s:1774
Wraps: (4) could not convert a single message before hitting truncation
Error types: (1) *withstack.withStack (2) *errutil.withPrefix (3) *withstack.withStack (4) *errutil.leafError

8

I’m seeing this same loop on the latest update. I originally updated because the previous version had a token-counting bug that resulted in a 7-day lockout, even on my Pro account.

While the changelog mentions a fix, the current build is now throwing agent executor error: could not convert a single message before hitting truncation for every model regardless of prompt size. The only exception is Claude 3.6 Sonnet, but even that reports “quota over” after just two small prompts.

1 Like
9

Trajectory ID: 0d077dd1-4131-4
Error: agent executor error: could not convert a single message before hitting truncation
(1) attached stack trace
– stack trace:
| google3/third_party/jetski/cortex/cortex.(*CascadeManager).executeHelper.func1
| third_party/jetski/cortex/cascade_manager.go:2096
| […repeated from below…]
Wraps: (2) agent executor error
Wraps: (3) attached stack trace
– stack trace:
| google3/third_party/jetski/cortex/chatconverters/chatconverters.init
| third_party/jetski/cortex/chatconverters/trajectory_chat_converter.go:44
| runtime.doInit1
| third_party/go/gc/src/runtime/proc.go:8105
| runtime.doInit
| third_party/go/gc/src/runtime/proc.go:8072
| runtime.main
| third_party/go/gc/src/runtime/proc.go:258
| runtime.goexit
| third_party/go/gc/src/runtime/asm_amd64.s:1774
Wraps: (4) could not convert a single message before hitting truncation
Error types: (1) *withstack.withStack (2) *errutil.withPrefix (3) *withstack.withStack (4) *errutil.leafError

I can’t even run a single command with this version 1.20.5 the model

10

Disable automatic updates and stick with 1.19.6 until the community stops complaining about the new releases.

1 Like
11

Downgrading to version 1.19.6 fixed my issue.

1 Like
12

Downgrading to version 1.19.6 fixed my issue as well.

1 Like
13

downgrade to old version fixed my issue as well, new version is bad

1 Like
14

lol, it’s been a month and still no update to fix what they broke. They’re really forcing people to stay on 1.19.6 because of their update

1 Like