inference-providers

Google Vertex AI

cloud_hosted

DocsConsole

Getting credentials

adc

adc via header (header Authorization: Bearer)

Environment: GOOGLE_APPLICATION_CREDENTIALS

gcloud auth application-default login; classic endpoints reject API keys (express mode adds API-key auth for some endpoints).

Credentials docs

Endpoints

IDURLProtocolAuth
generate-contenthttps://{region}-aiplatform.googleapis.com/v1/projects/{project}/locations/{region}/publishers/google/models/{model}:generateContentgoogle-generate-contentany

API surfaces

textstreamingembeddingsbatchcount_tokensprompt_caching

Quirks

No quirks recorded.

Models offered

ModelWire IDEndpointStatusReasoning
google/gemini-3-5-flash-litegemini-3.5-flash-litegenerate-contentgaeffort · default on · returns thought_parts · round-trips thought_signature
google/gemini-3-6-flashgemini-3.6-flashgenerate-contentgaeffort · default on · returns thought_parts · round-trips thought_signature
google/gemini-3-7-flashgemini-3.7-flashgenerate-contentgaeffort · default on · mandatory · returns thought_parts · round-trips thought_signature
google/gemini-3-progemini-3-pro-previewgenerate-contentprevieweffort · default on · mandatory · returns thought_parts · round-trips thought_signature