Google Vertex AI

Google Vertex AI is a separate provider from Google. Both call the same underlying Gemini models, but through different APIs: Google routes through the Gemini Developer API (generativelanguage.googleapis.com), while Google Vertex AI routes through Vertex AI Express Mode (aiplatform.googleapis.com) using a Vertex-issued API key. Pick this provider if your key was issued by Vertex AI rather than Google AI Studio.

Support tier

byok_only — LenserFight has a direct runtime path for this provider, but it is not wired to the Chainabit platform-credit gateway. You must supply your own API key.

Upstream docs

cloud.google.com/vertex-ai/generative-ai/docs

Models on LenserFight

Gemini 2.5 Pro (Vertex)

Field	Value
Key	gemini-2.5-pro-vertex
Capabilities	chat · reasoning · tools · vision
Context window	2 000 000 tokens
Input modalities	text · image · document
Output modalities	text
Provider docs	—

Gemini 2.5 Flash (Vertex)

Field	Value
Key	gemini-2.5-flash-vertex
Capabilities	chat · tools · vision
Context window	1 000 000 tokens
Input modalities	text · image · document
Output modalities	text
Provider docs	—

Usage notes

• Only text models are wired for this provider. Vertex-routed Imagen/Veo/Lyria variants are not registered — use the Google provider's Imagen 4 / Veo 3 / Lyria 2 for media generation.
• Configure your key with lf providers config google_vertex --from-env <YOUR_ENV_VAR> (reads from an environment variable instead of prompting, so the key never appears in shell history), then verify with lf providers test google_vertex.
• byok_only means this provider cannot be selected with platform_credit or sponsored funding — only user_byok_cloud.