Skip to content

feat(antigravity): add Gemini 3.8 Flash support - #5423

Open
Lynricsy wants to merge 2 commits into
router-for-me:devfrom
Lynricsy:feat/antigravity-gemini-3.8-flash
Open

feat(antigravity): add Gemini 3.8 Flash support#5423
Lynricsy wants to merge 2 commits into
router-for-me:devfrom
Lynricsy:feat/antigravity-gemini-3.8-flash

Conversation

@Lynricsy

@Lynricsy Lynricsy commented Sep 2, 2026

Copy link
Copy Markdown

Summary

  • register the official Antigravity gemini-3.8-flash-high and gemini-3.8-flash-medium routes
  • publish the 1,048,576-token context limit, 65,536-token output limit, and supported text/image/audio/video inputs for both routes
  • expose exactly the supported low, medium, and high thinking levels; Gemini 3.8 Flash rejects minimal
  • add registry contract tests covering both route identities and capabilities

Source distinction

  • CLIProxyAPI's canonical remote catalog already contains gemini-3.8-flash-high; this change synchronizes that metadata into the embedded catalog.
  • Google Antigravity's Headless documentation additionally lists gemini-3.8-flash-medium. The paired models catalog PR #50 adds that official slug to the remote source so startup and periodic catalog refreshes preserve it.

First-party model facts

  • Google Antigravity lists Gemini 3.8 Flash with Low, Medium, and High reasoning levels.
  • Google's Sep 1 launch post lists pricing at $0.75 per million input tokens and $3.75 per million output tokens.

Sources

Verification

  • go test ./internal/registry
  • go test -run TestRegisterModelsForAuth_AntigravityFetchesWebSearchCapability ./sdk/cliproxy
  • go build -o test-output ./cmd/server && rm test-output

Related to #5421 (currently open).

Co-authored-by: Wine Fox <fox@ling.plus>
Copilot AI lite review requested due to automatic review settings September 2, 2026 16:12
@github-actions
github-actions Bot changed the base branch from main to dev September 2, 2026 16:12
@github-actions

github-actions Bot commented Sep 2, 2026

Copy link
Copy Markdown

This pull request targeted main.

The base branch has been automatically changed to dev.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Sep 2, 2026

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review Completed 2026-09-02T16:20:24.588025Z 250c9ea New commits
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🟡 Changes recommended

The PR description claims minimal is rejected, but existing thinking validation can clamp unsupported levels in some request paths, so the documented behavior should be aligned with actual behavior or the validation adjusted.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

Adds Gemini 3.8 Flash support to the static Antigravity model catalog, publishing its token limits, supported modalities, and thinking-level capabilities, plus a registry contract test to ensure the metadata remains stable.

Changes:

  • Added a new Antigravity model definition for gemini-3.8-flash-high with 1,048,576 context / 65,536 output token limits and text/image/audio/video inputs.
  • Restricted advertised thinking levels to low/medium/high for Gemini 3.8 Flash in the static catalog.
  • Added a contract test validating the model identity and capabilities.
File summaries
File Description
internal/registry/models/models.json Registers the gemini-3.8-flash-high Antigravity model with token limits, thinking levels, and supported modalities.
internal/registry/model_definitions_test.go Adds a contract test asserting the new model is present and its metadata matches expectations.
Review details
  • Files reviewed: 2/2 changed files
  • Comments generated: 1
  • Review effort level: Lite

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

Comment on lines +67 to +72
if got.Thinking == nil || got.Thinking.Min != 1 || got.Thinking.Max != 65535 || !got.Thinking.DynamicAllowed {
t.Fatalf("Gemini 3.8 Flash thinking metadata = %+v", got.Thinking)
}
if want := []string{"low", "medium", "high"}; !slices.Equal(got.Thinking.Levels, want) {
t.Fatalf("Gemini 3.8 Flash thinking levels = %v, want %v", got.Thinking.Levels, want)
}
Co-authored-by: Wine Fox <fox@ling.plus>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Please add support for Gemini 3.8 Flash on the Antigravity platform as soon as possible. Thank you!

2 participants