Voyage 3.5

Overview

Voyage 3.5 series comprises two models designed for enterprise-grade semantic search and AI applications, delivering state-of-the-art performance on industry benchmarks.

Models

Voyage 3.5

• High-performance flagship model
• Optimized for accuracy and semantic understanding
• Suitable for production deployments
• Price: $0.06 per 1 million tokens

Voyage 3.5 Lite

• Lightweight variant for cost-sensitive applications
• Faster inference with maintained quality
• Ideal for high-volume scenarios
• Price: $0.02 per 1 million tokens

Voyage 4 Series (Latest)

The Voyage 4 series introduces an industry-first capability: shared embedding spaces, allowing models to work together seamlessly.

voyage-4-large surpasses competing models:

• 3.87% better than Gemini Embedding 001
• 8.20% better than Cohere Embed v4
• 14.05% better than OpenAI v3 Large

Features

• State-of-the-art retrieval performance
• Multilingual support
• Long context handling
• Domain-specific fine-tuning available
• Enterprise-grade reliability
• Flexible API integration
• Batch processing support

Use Cases

• Enterprise semantic search
• Retrieval-Augmented Generation (RAG)
• Document similarity and clustering
• Question answering systems
• Cross-lingual information retrieval
• Recommendation engines

API Integration

Simple REST API with client libraries for:

• Python
• JavaScript/TypeScript
• Java
• Go

Performance

Achieves top scores on MTEB (Massive Text Embedding Benchmark) and other industry-standard benchmarks, particularly excelling in retrieval tasks.

Pricing

• Voyage 3.5: $0.06 per 1M tokens
• Voyage 3.5 Lite: $0.02 per 1M tokens
• Voyage 4 Large: $0.12 per 1M tokens
• Free tier available for development and testing