voyage-code-2
Voyage Code 2 specializes in finding code from text or related code. Its release describes contrastive training on large code-pair collections.
Model design
The launch evaluates 11 code retrieval tasks, including Python, C++ and Java material. That evaluation scope is narrower than a claim to understand every language.voyage-code-2 code retrieval
Practical use
Practical guidance. Keep this release identifiable when maintaining an older code index. Moving to a later Code model changes the vector representation. Re-embed a representative repository and check whether searches still find the correct function and surrounding implementation.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | API |
| Atlas lifecycle | Legacy since 15 January 2026MongoDB Voyage AI model lifecycle: Atlas scope and release history |
| Input and output | 16,000 tokens; 1,536 dimensionsMongoDB Voyage AI model specifications |
Related models
- voyage-code-4
Voyage AI · Voyage Code
Code embeddings trained with completed pull requests and repository tasks.
Launch date API - voyage-code-3
Voyage AI · Voyage Code
Code embeddings with variable dimensions and quantization support.
Launch date API