feat(ai): add Gemini 3.1 Pro support for google-antigravity provider

[yorha59]

Summary

Add Gemini 3.1 Pro model support for the google-antigravity provider.

Google has deprecated Gemini 3 Pro (the sandbox now returns "Gemini 3 Pro is no longer available. Please switch to Gemini 3.1 Pro"), so this PR adds the new model variants.

Changes

1. packages/ai/scripts/generate-models.ts

• Added gemini-3.1-pro-high and gemini-3.1-pro-low to the antigravityModels array

2. packages/ai/src/models.generated.ts

• Added corresponding model entries for gemini-3.1-pro-high and gemini-3.1-pro-low under the google-antigravity provider

3. packages/ai/src/providers/google-gemini-cli.ts

• Updated DEFAULT_ANTIGRAVITY_VERSION from 1.15.8 to 1.18.0 (required — the sandbox returns "not available on this version. Please upgrade" for UA < 1.18.0)

Testing

• Verified both gemini-3.1-pro-high and gemini-3.1-pro-low return 200 OK from the antigravity sandbox API
• Tested via OpenClaw with real conversations — reasoning and tool use work correctly
• Confirmed older models (gemini-3-pro-high, gemini-3-flash) still work with the new UA version

Notes

• Model IDs follow the existing naming convention: gemini-3.1-pro-high (deeper reasoning) and gemini-3.1-pro-low (faster, lighter reasoning)
• The UA version bump to 1.18.0 is required by Google's sandbox — versions below 1.18.0 are rejected for 3.1 models
• gemini-3.1-pro (without suffix) returns 404; the -high/-low suffix is mandatory

[github-actions]

Hi @yorha59, thanks for your interest in contributing!

We ask new contributors to open an issue first before submitting a PR. This helps us discuss the approach and avoid wasted effort.

Next steps:

1. Open an issue describing what you want to change and why (keep it concise, write in your human voice, AI slop will be closed)
2. Once a maintainer approves with lgtm, you'll be added to the approved contributors list
3. Then you can submit your PR

This PR will be closed automatically. See https://github.com/badlogic/pi-mono/blob/main/CONTRIBUTING.md for more details.

[51616]

I was gonna post a PR but I guess I'm an hour late lol.

I know I'm not the maintainer but a nice UX improvement would be to make the thinking options for -low to be only {off,low} and -high to be only {off,high}.