Lowest price
Choose text-embedding-3-small
text-embedding-3-small costs $0.02 / 1M tokens, against $0.13 / 1M tokens for text-embedding-3-large.
text-embedding-3-large vs text-embedding-3-small: price, context and capabilities
Choose text-embedding-3-small for the lowest price; on the other points the two are level or the data is missing.
In short
- text-embedding-3-small is 85% cheaper than text-embedding-3-large at the headline price ($0.02 / 1M tokens vs $0.13 / 1M tokens).
- Both accept 8k tokens per request.
Which should you choose?
Each answer below comes straight from the published prices, specifications and independent benchmark results.
Specifications side by side
| text-embedding-3-large | text-embedding-3-small | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Category | Embeddings | Embeddings |
| Status | Generally available | Generally available |
| Headline price | $0.13 / 1M tokens | $0.02 / 1M tokens |
| Context window | 8k | 8k |
| Max output | No data | No data |
| Released | January 2024 | January 2024 |
| Knowledge cutoff | No data | No data |
| Licence | Proprietary | Proprietary |
| Capabilities | No data | No data |
| API model id | text-embedding-3-large | text-embedding-3-small |
Frequently asked questions
Should I choose text-embedding-3-large or text-embedding-3-small?
Choose text-embedding-3-small for the lowest price; on the other points the two are level or the data is missing.
Which is cheaper, text-embedding-3-large or text-embedding-3-small?
text-embedding-3-small is 85% cheaper than text-embedding-3-large at the headline price ($0.02 / 1M tokens vs $0.13 / 1M tokens).
Which has the larger context window, text-embedding-3-large or text-embedding-3-small?
Both accept 8k tokens per request.
What is the difference between text-embedding-3-large and text-embedding-3-small?
text-embedding-3-large is made by OpenAI and text-embedding-3-small by OpenAI. text-embedding-3-small is 85% cheaper than text-embedding-3-large at the headline price ($0.02 / 1M tokens vs $0.13 / 1M tokens). Both accept 8k tokens per request.