Updated

Embed 5 Pro and Embed 5 Fast

Cohere released Embed 5, a family of multimodal embedding models in Pro and Fast tiers. Both support text and image inputs, more than 100 languages, and a 128K-token context; their shared embedding space lets teams index with Pro and query with either tier.

By No Code MBA · Facts checked

Provider
Cohere
Announcement date
Status at last check
Generally available
Model family
Embed 5
Preceding model
Embed 4
Capabilities
Embeddings, Text, Image

What changed

Embed 5 comes in Pro and Fast tiers that share one embedding space, so a corpus indexed with Pro can be queried with either. Both take text, images or fused text and image, support 100+ languages and a 128K-token context, and output Matryoshka embeddings from 256 to 2048 dimensions in float, int8 or binary. Cohere reports a ViDoRe V3 average of 85.8 for Pro and 84.5 for Fast versus 77.0 for Embed 4, and says Fast processes about 2.4x more documents per second than Pro.

Access and availability

Use the Cohere API with embed-v5.0-pro or embed-v5.0-fast, including batch jobs via the Embed Jobs API. Both are also on Microsoft Foundry, Amazon SageMaker, Cohere North, and private VPC or on-premises deployments with vLLM.

Official access documentation

Pricing

Pro: $0.12 per million text tokens and $0.40 per million image tokens. Fast: $0.08 per million text tokens and $0.40 per million image tokens

Official pricing

What you can build

Test both tiers on representative text, image-rich documents, and multilingual queries; compare retrieval quality, latency, and index consistency before choosing production roles.

Limitations

Benchmarks are self-reported and use Cohere's own RCP-nDCG@10 method, which measures reranking over a fixed candidate set. Cohere's tables show Gemini Embedding 2 ahead on several languages, binary vectors trade some accuracy, and cross-tier search needs matching dimensions.

Sources and release history