text-embedding-3-large

Cost & access

text-embedding-3-large is priced per million embedding tokens (or per request), not like chat completion. Compare dimensions and batch APIs on OpenAI's pricing page. With a 128K tokens context window, long PDFs or chat histories increase input tokens quickly—trim history or summarize older turns in production.

Functional understanding

• OpenAI embedding model for semantic search.
• Modalities: text · License: Proprietary · Released 2024-01-25.
• Best-fit workflows for this model:
• • Semantic search, deduplication, and RAG retrieval—text-embedding-3-large outputs vectors, not chat prose.
• • Clustering support tickets, docs, or product catalogs by meaning.

Technical foundation

• OpenAI reports Undisclosed parameters; training data: Undisclosed.
• Context: 128K tokens. Open weights: no.
• text-embedding-3-large is positioned as a embedding model in the OpenAI lineup.

First API call

Use OpenAI's embeddings endpoint; pass text-embedding-3-large as the model name and store vectors in your DB or vector index.

from openai import OpenAI

client = OpenAI()
vec = client.embeddings.create(
    model="text-embedding-3-large",
    input="Text to embed for text-embedding-3-large",
)
print(len(vec.data[0].embedding), "dimensions")

Important technical topics

• Prompting text-embedding-3-large: be explicit about output format. Weak: "Analyze this." Better: "Return JSON with fields id, total, date for OpenAI billing data."
• Temperature: use 0–0.3 for extraction and compliance on text-embedding-3-large; 0.7–1.0 for brainstorming.
• Tokens: text-embedding-3-large bills by tokens (~¾ word each). Undisclosed parameters affect capability; your bill is driven by context length and call volume.
• Context window (128K tokens): everything in one request—system prompt, tools, RAG chunks, and history—must fit. Truncate or summarize when approaching the limit for text-embedding-3-large.

Real enterprise patterns

• Index docs with text-embedding-3-large, then retrieve top-k chunks before calling a chat model.
• Hybrid search: combine keyword (BM25) + embeddings for better recall.
• Version embedding indexes when you change models—dimensions may differ.
• Monitor drift: re-embed when OpenAI ships a new embedding revision.

Production & security

• Secrets: never commit keys for text-embedding-3-large; use vault + per-environment rotation.
• PII: mask before inference; log redacted prompts only.
• Observability: trace id per request; log model=text-embedding-3-large, tokens in/out, latency.
• Rate limits: handle OpenAI 429/5xx with exponential backoff and circuit breakers.
• Guardrails: schema-validate JSON; block disallowed topics; cross-check numbers against source docs.

Mini projects with this model

• Semantic FAQ: embed help-center articles with text-embedding-3-large, answer from nearest neighbors.
• Duplicate ticket detector for support queues.
• Product recommendation by description similarity.
• Eval: NDCG@k on labeled query–doc pairs.

Suggested stack

• Language: Python 3.11+
• Embeddings: text-embedding-3-large (OpenAI)
• Vector DB: Pinecone, Chroma, pgvector, or Weaviate
• Orchestration: LangChain or LlamaIndex for chunking
• UI: Streamlit or Next.js for internal tools
• APIs: FastAPI

Learning path

• Python basics
• HTTP/REST and environment variables
• OpenAI embeddings API for text-embedding-3-large
• Vector database fundamentals
• Chunking strategies
• RAG retrieval evaluation
• Deploy index refresh pipeline