Skip to content
AnyRouter

Gemini Embedding 2

Google
Compare

Gemini Embedding 2 is Google's first multimodal embedding model, mapping text, images, video, audio, and PDFs into one 3,072-dimension vector space for cross-modal semantic search, document retrieval, and recommendations over 100+ languages. Upstream accepts 8,192 input tokens and MRL-truncates to 128–3,072 dimensions. AnyRouter exposes the text path only — the OpenAI-compatible /embeddings body is a text `input`, so the extra upstream modalities are deliberately not declared here.

Providers
Capabilities
OpenRouter
openrouter-byok
$0
$0
Unavailable
Usage analytics

Loading usage…

API & code

Embeddings API

/api/v1/embeddings

Generate embedding vectors for search, RAG, and clustering.

Uptime & Health
No uptime data yet

These providers haven't been health-probed for this model yet. The router still routes around upstreams that fail live requests — uptime fills in once probe history accrues.

Share cards
Gemini Embedding 2 share card
Gemini Embedding 2
OpenRouter upstream share card
OpenRouter upstream
Credits
Use your own key

Run Gemini Embedding 2 on your own key — your requests are billed by the provider. Pool callers pay AnyRouter credits.

No BYOK keys configured for this model yet.

Share a key with the pool to earn credits for every request it serves, covering your plan cost.

Embedding vectors
Vector dimensions3,072
Max input8,192 tokens
Price$0.2 / 1M tokens
Request parameters
inputmodeldimensionsencoding_format
ArchitectureTransformer
Categoryembedding
ReleasedApr 22, 2026
Modalities
→
Capabilities
Embeddings are fixed-length vectors — compare them with cosine similarity for semantic search, RAG retrieval, clustering, and deduplication. Embed queries and documents with the same model, or the distances are meaningless.
Back to models