You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add the Gemini image generation models under google-vertex, now that
they're GA (and 2.5 Flash Image as it's non-preview was also not yet present Vertex):
gemini-2.5-flash-image (Nano Banana)
gemini-3.1-flash-image (Nano Banana 2)
gemini-3-pro-image (Nano Banana Pro)
Each inherits provider-agnostic facts via base_model from models/google/
and declares only the Vertex-specific cost and reasoning_options,
following the existing gemini-3.5-flash.toml pattern.
Pricing/limits verified against Vertex AI docs. Note output uses the
image-output token rate (repo convention for these models). Context for
3 Pro Image is 65,536/32,768 on Vertex (no 131K override).
AUTOMATED REVIEW: I found a few data correctness issues that should be fixed before merge.
providers/google-vertex/models/gemini-3.1-flash-image.toml:1 inherits preview limits from google/gemini-3.1-flash-image-preview. The resolved Vertex entry becomes context = 65_536 and output = 65_536, but current Vertex docs list stable gemini-3.1-flash-image as context = 131_072 and output = 32_768. Please add a [limit] override.
providers/google-vertex/models/gemini-2.5-flash-image.toml:1-2 publishes reasoning = true via the base model plus reasoning_options = [], but Vertex/Gemini docs list gemini-2.5-flash-image as Thinking not supported. Please override reasoning = false and remove reasoning_options for this provider entry.
providers/google-vertex/models/gemini-3-pro-image.toml:1 leaves structured_output unset via the preview base, but current docs list stable gemini-3-pro-image structured outputs as supported. Please add structured_output = true.
providers/google-vertex/models/gemini-3-pro-image.toml:1 and providers/google-vertex/models/gemini-3.1-flash-image.toml:1 inherit preview release_date / last_updated values under stable GA IDs. Vertex docs list both stable IDs as GA with release date 2026-05-28; these should be overridden or represented by non-preview base metadata.
Checks run locally: bun validate passed, and git diff --check dev...HEAD passed.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Add the Gemini image generation models under
google-vertex, now thatthey're GA (and 2.5 Flash Image as it's non-preview was also not yet present Vertex):
gemini-2.5-flash-image(Nano Banana)gemini-3.1-flash-image(Nano Banana 2)gemini-3-pro-image(Nano Banana Pro)Each inherits provider-agnostic facts via
base_modelfrommodels/google/and declares only the Vertex-specific
costandreasoning_options,following the existing
gemini-3.5-flash.tomlpattern.Pricing/limits verified against Vertex AI docs. Note
outputuses theimage-output token rate (repo convention for these models). Context for
3 Pro Image is 65,536/32,768 on Vertex (no 131K override).