t

text-embedding-3-large

Model spec
Provider
OpenAI
Context window
8K ctx
Max output
-
Input price
$0.13 / 1M tokens
Output price
-
License
Proprietary
Open weights
No

text-embedding-3-large is OpenAI’s premium text embedding model, optimized for converting text into numerical representations suitable for machine learning applications. It excels in tasks requiring deep semantic understanding, such as search and content organization, where accuracy and nuanced representation are critical.

The model fits within OpenAI’s lineup as their most advanced embedding option, positioned above smaller or more specialized alternatives. Its large context window allows it to handle substantial documents without truncation, a key differentiator for processing complex texts.

Developers building enterprise-grade search systems or retrieval pipelines should consider this model, particularly if they prioritize performance and have budget for proprietary solutions. Those requiring open-weight models or transparent pricing structures may need to evaluate alternatives.

Modality

Text

Use cases

Semantic search applicationsDocument clustering and categorizationRetrieval-augmented generation (RAG) systemsRecommendation systems based on text similarityDuplicate detection across text corpora

Pros & cons

Pros

  • High-quality embeddings optimized for search and retrieval tasks
  • Supports a context window of 8191 tokens, enabling processing of longer documents
  • Proven integration with OpenAI's ecosystem and tools

Cons

  • Proprietary license restricts modification or redistribution
  • No open-weight availability limits transparency and customization
  • Input pricing may be prohibitive for high-volume use cases
  • Output capabilities and pricing are not specified here

Related models

E
Embedding

512 ctx · $0.10 in per 1M · Cohere

Cohere's Embed v3 is a text embedding model designed for semantic search and retrieval tasks, with support for multiple ...

View details Visit

Read the official docs

text-embedding-3-large

View documentation