Not published
Only provider-documented reasoning controls are shown; missing controls remain unpublished rather than inferred.
MODEL REFERENCE / GOOGLE
Gemini Embedding 2
verified model reference.
Google stable multimodal embedding model for semantic search and cross-modal retrieval across text, images, audio, video and PDFs.
8,192context tokens
Not publishedmax output
2026-10-12last verified
PRIMARY-SOURCE VERIFIED
Gemini Embedding 2 at a glance.
Google documents a 8,192-token context window. Input modalities: text + image + video + audio + PDF. Output: embedding. Pricing status: Official paid API.
MODEL IDgemini-embedding-2
CONTEXT WINDOW8,192tokens
MAX OUTPUTNot published
KNOWLEDGE CUTOFFNot published
STANDARD INPUT$0.20per 1 million TEXT input tokens (other media separate)
STANDARD OUTPUT—per 1 million TEXT input tokens (other media separate)
CAPABILITY CONTRACT
Published model behavior.
text + image + video + audio + PDF
Input modalities come from the cited official model documentation.
embedding
Output modality and output-limit claims remain separate so an undocumented token cap is not guessed.
PRICING STATUS
Official paid API.
$0.20 input per 1 million TEXT input tokens (other media separate).
SXF preserves the provider-published native billing basis. Only calculator-eligible generative token models enter token-cost arithmetic; specialist units and input-only embeddings remain visible without forced conversion.
CAVEATS
What the source record does not assume.
01
Displayed Standard price ($0.20 per 1M tokens) applies to TEXT embedding inputs only. Embeddings have no generative output-token charge.
02
Google separately bills IMAGE input at $0.45 per 1M image tokens ($0.00012/image), AUDIO at $6.50 per 1M audio tokens ($0.00016/second), and VIDEO at $12.00 per 1M video tokens ($0.00079/frame).
03
Batch pricing differs: $0.10 text, $0.225 image, $3.25 audio and $6.00 video per 1M corresponding input tokens.
04
Output embedding dimensions are configurable 128–3072 (recommended 768, 1536 or 3072). Embedding 2 vectors are NOT directly compatible with Gemini Embedding 001 vectors; migration requires re-embedding existing documents.
05
The generic generative text-cost calculator excludes input-only embedding inference.
EXPLORE
Keep the model in context.
OFFICIAL SOURCES
Verification links.
WHAT CHANGED
Gemini Embedding 2: what changed.
No post-baseline factual changes recorded for Gemini Embedding 2 since 2026-10-12.
VERIFIED BASELINE2026-10-12Current facts remain aligned with the SXF ledger.
VERIFIED CHANGE HISTORY
Gemini Embedding 2 verified history.
Gemini Embedding 2
Baseline verified: 8,192 context · Not published max output · $0.20 input per 1 million TEXT input tokens (other media separate)
Official evidence ↗Append-only SXF ledger · 1 verified event · chain head b12eeb63a07e…