C

Cohere: Embed 4

Cohere
Vision
128K
Context
$0.1200
Input /1M
N/A
Output /1M
N/A
Max Output
Run API Call Open in Chat

Cohere: Embed 4

Model Overview

Property Value
Model ID cohere/embed-v4
Provider Cohere
Category Embedding
Released April 15, 2025
Status Active
Flagship Yes

Description

Embed 4 is Cohere's most performant multilingual multimodal embedding model. It transforms different modalities such as images, texts, and interleaved images and texts into a single vector representation. Embed 4 offers state-of-the-art performance in Text-to-Text, Text-to-Image, and Text-to-Mixed Modality domains across 100+ languages. It can vectorize interleaved texts and images and capture key visual features from screenshots of PDFs, slides, tables, figures, and more. Output is a 1,536-dimensional vector with support for Matryoshka dimensions [256, 512, 1024, 1536].

Specifications

Property Value
Context Window 128,000 tokens
Max Output Tokens 1,536 dimensions
Knowledge Cutoff N/A
Modalities Text, Image, Mixed
Speed Tier Fast
Quality Tier Premium

Pricing

Type Price
Text Input $0.12 per 1M tokens
Image Input $0.47 per 1M image tokens

Capabilities

Capability Supported
Vision/Image Input Yes
Tool/Function Calling No
JSON Mode No
Streaming No
System Prompt No

Use Cases

  • Semantic search and retrieval
  • Document similarity comparison
  • Clustering and categorization
  • Recommendation systems
  • Vector-based semantic matching

Strengths

  • Very large context window (128K) for long-form content
  • Multimodal support (text, images, mixed)
  • 100+ languages supported
  • Matryoshka embeddings for flexible compression
  • State-of-the-art retrieval performance

Limitations

  • Embedding model, not suitable for text generation
  • Image embeddings are more expensive than text
  • Specialized model not suitable for general chat

Integration with LangMart

Gateway Support:

  • Type 1 (Full Platform): Yes
  • Type 2 (Cloud Gateway): Yes
  • Type 3 (Self-hosted): Yes

API Endpoint:

curl -X POST https://api.langmart.ai/v1/embeddings \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "cohere/embed-v4",
    "input": ["Hello world", "How are you?"],
    "input_type": "search_document"
  }'

OpenAI Compatibility: Full compatibility with OpenAI API format.

Models in this category and from this provider are available in the LangMart marketplace.

Additional Resources

Data Source: LangMart Provider Models Database Last Updated: 2026-01-15