Voyage 3
Voyage AIEmbedding Models
92.5
Performance
★ 4.6
Rating
310
Reviews
EmbeddingProprietary
About
Top-performing embedding model on MTEB with optimized retrieval quality for RAG applications and long-context support.
Strengths
Leading MTEB scores among embedding models. 1,024 dimensions with 32K context window. Excellent for code and general-purpose retrieval. Optimized specifically for RAG applications. Low-cost at $0.06/1M tokens.
Specifications
- Context window
- 32,000
- Parameters
- —
Pricing
- Input cost
- $0.06 / 1M
- Output cost
- —
$0.06 per 1M tokens. Free tier available (50M tokens/mo).
Available On
Voyage AI API
Features
text embeddingcode embeddingbatch processinglong context
Performance Trend
Benchmark score trends over time for the top 5 benchmarks.
Loading history...
Benchmarks
Scores from various benchmark tests; higher is better.
| Test | Score | Percentile | Source |
|---|---|---|---|
| MTEB | 67.5 | p97 | seed |
