Model comparison
Compare text-embedding-3-large and Voyage Code 4 using the same provider-sourced embeddings & vector search rubric. No mystery score and no invented benchmark ranking.
Facts checked September 4, 2026
Set your usage. Your estimate updates as you type.
Assumes 600 tokens per page, processed separately. Actual token counts vary. This covers embedding only, not storage, search, or generated answers.
Estimates exclude taxes, tools, cache storage/writes, free allowances and custom discounts. Image estimates cover output only, not prompt or reference-image charges. Quality modes differ by model. Unlisted settings are not treated as free.
| Model | Access | Estimated total (USD) |
|---|---|---|
| Voyage Code 4Voyage AI | Voyage AI | $0.0007 |
| text-embedding-3-largeOpenAI | OpenAI | $0.0008 |
Quick take
OpenAI's highest-capability text embedding model for semantic search, recommendations, clustering, and retrieval pipelines.
The model embeds text only and does not chunk long documents for you. Your ingestion pipeline still needs a deliberate chunking, metadata, evaluation, and re-indexing strategy.
Voyage AI's specialist embedding model for finding relevant code from natural-language questions or other source-code context.
A specialist code model is not automatically better for README files, tickets, or business documentation. Evaluate mixed repositories separately and account for recurring re-index costs.
Compare the published facts
Values use each provider's own published units and limits. A blank means the provider did not publish a directly comparable value in the sources reviewed.
| Embeddings & vector search | text-embedding-3-large | Voyage Code 4 |
|---|---|---|
| Embedding priceCurrent provider price per million input tokens or the closest published billing unit. | $0.13 / 1M tokens | $0.12 / 1M tokens after 200M free |
| Input capacityMaximum content accepted in one embedding input, using the provider's documented token basis. | 8,191 input tokens | 32K tokens |
| Vector dimensionsSupported output sizes; smaller vectors reduce storage while larger vectors may preserve more information. | 3,072 default; shorter vectors via dimensions | 1,024 default; 256, 512, or 2,048 optional |
| Accepted inputsText, code, image, audio, video, PDF, or document-aware input supported by the endpoint. | Text | Code and natural-language text |
| Retrieval controlsQuery/document modes, task types, truncation, chunking, or other controls that shape vectors for retrieval. | Optional dimensions parameter; chunking handled by the application | Query/document modes, truncation, and shared Voyage 4 vector space |
| Where it runsDirect API, cloud marketplace, private deployment, or self-hosted route documented by the provider. | OpenAI hosted Embeddings API | Voyage API and MongoDB-integrated routes |
How to choose
Start with the job you need to complete, then validate cost, access, and policy details on your exact provider route.
OpenAI says business and API data is not used to train its models by default. Abuse-monitoring retention and eligible zero-data-retention controls depend on the endpoint and organization approval.
Provider and API links
Voyage's public terms allow customer content to improve services unless the customer opts out. Eligible paid organizations can configure an opt-out and zero-day retention; separately negotiated enterprise terms may differ.
Provider and API links
Frequently asked questions
text-embedding-3-large: OpenAI's highest-capability text embedding model for semantic search, recommendations, clustering, and retrieval pipelines. Voyage Code 4: Voyage AI's specialist embedding model for finding relevant code from natural-language questions or other source-code context.
Consider text-embedding-3-large when your priority is High-quality text search and RAG. Consider Voyage Code 4 when your priority is Repository and symbol search. Test both with your own data and provider route before committing.
No. This comparison aligns provider-published facts for the Embeddings & vector search category. It does not claim a universal winner or combine incompatible third-party benchmark scores.
