Skip to main content

Embeddings:Built into Weaviate Cloud

Create embeddings in Weaviate Cloud without managing multiple embedding providers.

Weaviate Embeddings architecture

Fast, flexible development

Simplify operations with one less API and vendor to manage. Choose from class-leading open-source models.

Models close to your data

Bring models closer to your data.

GPU-powered and cost-efficient

Maximize performance while managing costs with simple, pay-as-you-go pricing.

No external provider needed

Access embedding models directly in Weaviate Cloud without sending data to a third party.

What is Weaviate Embeddings?

Embeddings, right next to your data

Weaviate Embeddings is a service in Weaviate Cloud that simplifies the creation and management of vector embeddings.

Weaviate Embeddings architecture

Access leading models without leaving Weaviate

Developers can use a range of embedding models without needing to send data to an external provider:

  • ✓One less API and vendor to manage
  • ✓Class-leading open-source models
  • ✓Lower latency, closer to your data
  • ✓Simple pay-as-you-go pricing
Read the blog post to learn more

Available models

Snowflake Arctic L v2.0 (Default)

Multilingual support and up to 8,192 tokens. Best for big-scale, complex retrieval.

Snowflake Arctic M v1.5

English support and up to 512 tokens. Best for fast, lightweight retrieval.

ModernVBERT ColModernVBERT

Multimodal, English support for images and document pages (query text limit: 8,092 tokens). Best for visual documents like PDFs, slides and invoices without OCR preprocessing.

Pricing

Simple, pay-as-you-go pricing

Pay only for the tokens you embed.

Snowflake Arctic Embed 1.5

Text embedding

$0.025 per 1M tokens

Snowflake Arctic Embed 2.0

Text embedding

$0.040 per 1M tokens

ModernVBERT ColModernVBERT

Multimodal embedding

$0.065 per 1M tokens

Generate embeddings
inside Weaviate Cloud

Open in Weaviate Cloud