The Text Embedding 3 Large API is the retrieval layer under our RAG stack — top recall, and the 5% RouterBase discount is pure margin at our embedding volume.
The Text Embedding 3 Large API is OpenAI's highest-quality embedding model — it turns text into dense vectors of up to 3,072 dimensions, accepts up to 8,191 tokens per input, and lets you shrink the output via the `dimensions` parameter to trade vector size against cost. The Text Embedding 3 Large API is OpenAI-compatible: point any existing SDK at RouterBase and set the model to text-embedding-3-large. Billed at $0.1235 / 1M input tokens — 5% below the standard rate — through one REST endpoint.
Output
SampleRun the model to replace this previewText Embedding 3 Large API: Embeddings
Use the Text Embedding 3 Large API to turn text into OpenAI's highest-quality embeddings — up to 3072 dimensions, an 8191-token input, and adjustable output size.
The Text Embedding 3 Large API is OpenAI's most capable embedding model — it turns text into dense vectors of up to 3,072 dimensions, accepts up to 8,191 tokens per input, and lets you shrink the output via the `dimensions` parameter to trade vector size against cost and accuracy. Routed through RouterBase, the Text Embedding 3 Large API is fully OpenAI-compatible: point any existing SDK at RouterBase's base URL, set the model to text-embedding-3-large, and your first Text Embedding 3 Large API call returns vectors immediately.
Pricing for the Text Embedding 3 Large API is $0.1235 / 1M input tokens — 5% below the official published rate. One RouterBase key — no OpenAI account required.
Six reasons teams build retrieval on the Text Embedding 3 Large API
From 3072-dimensional vectors to 5%-off pricing — what makes the Text Embedding 3 Large API stand out.
Highest-quality OpenAI embeddings
The Text Embedding 3 Large API produces OpenAI’s most accurate embeddings — up to 3,072 dimensions for top retrieval and ranking quality across languages and domains.
Adjustable dimensions
Pass the `dimensions` parameter and the Text Embedding 3 Large API truncates the vector — shrink to 1024 or 256 dims to cut storage and speed up search with minimal accuracy loss.
8191-token input
Each Text Embedding 3 Large API request embeds up to 8,191 tokens — long documents, code files, or full chunks fit in a single call without extra splitting.
Built for RAG & semantic search
The Text Embedding 3 Large API powers retrieval-augmented generation, semantic search, clustering, classification, deduplication, and recommendations — the backbone of modern RAG stacks.
OpenAI-compatible endpoint
The Text Embedding 3 Large API speaks the OpenAI /v1/embeddings wire format. Point any OpenAI SDK at RouterBase — no separate credentials or call-site changes needed.
One key for 200+ models
The same RouterBase key that calls the Text Embedding 3 Large API also routes to GPT-5.4, Claude Opus 4.8, Gemini 3 Pro, and 200+ other models — no per-provider credential management.
Quickstart
Get started with the Text Embedding 3 Large API in 3 steps
From sign-up to your first vector in under 5 minutes.
Create a RouterBase API key
Sign up and generate an API key — one key reaches the Text Embedding 3 Large API and every other model in the catalog.
Request an embedding
POST to routerbase.com/v1/embeddings with model text-embedding-3-large and your input text (or an array of texts). The Text Embedding 3 Large API returns vectors in the standard OpenAI embeddings schema.
Store & search
Write the vectors to your vector database and query by cosine similarity. Every Text Embedding 3 Large API response includes a token-usage breakdown so cost is visible on each call.
Pay only for what you use
RouterBase passes through partner-tier pricing. Compared against the model's official published API rate.
What teams build with the Text Embedding 3 Large API
Real production retrieval running on the Text Embedding 3 Large API and the RouterBase model catalog.
We truncate to 1024 dims via the Text Embedding 3 Large API and cut our vector storage in half with almost no drop in search quality.
One array request to the Text Embedding 3 Large API embeds a whole batch of chunks. Re-indexing our corpus got dramatically simpler.
Semantic search on the Text Embedding 3 Large API beat our old model on recall. RouterBase makes it 5% cheaper and routes around outages automatically.
8191 tokens per input means I embed whole docs without fiddly splitting. The Text Embedding 3 Large API just handles them.
RouterBase puts the Text Embedding 3 Large API and 200+ other models behind one key. Our team stopped filing requests for new provider accounts.
Same key, one model field — we embed with the Text Embedding 3 Large API and generate with GPT-5.4. The whole RAG pipeline runs through RouterBase.
We moved our semantic search to the Text Embedding 3 Large API over a weekend. The only PR comment was 'wait, that's all?'.
At $0.1235 / 1M tokens minus 5%, the Text Embedding 3 Large API gave us best-in-class retrieval at a price that scales to millions of documents.
Frequently Asked Questions
Common questions about the Text Embedding 3 Large API.
It is RouterBase's pass-through to OpenAI's text-embedding-3-large — OpenAI's highest-quality embedding model, turning text into dense vectors of up to 3,072 dimensions, served via an OpenAI-compatible /v1/embeddings REST interface.