POST https://studio.dataeyes.ai/v1/embeddings| Model | Default dimensions | Supports dimensions |
|---|---|---|
text-embedding-3-large | 3072 | Yes |
text-embedding-3-small | 1536 | Yes |
text-embedding-ada-002 | 1536 | No |
{
"object": "list",
"model": "text-embedding-3-large",
"data": [
{ "object": "embedding", "index": 0, "embedding": [-0.08599853515625, 0.05499267578125, -0.07855224609375, "…256 in total"] },
{ "object": "embedding", "index": 1, "embedding": [0.04443359375, -0.00901031494140625, -0.006526947021484375, "…256 in total"] }
],
"usage": { "prompt_tokens": 15, "total_tokens": 15 }
}| Parameter | Type | Required | Description |
|---|---|---|---|
model | string | Yes | Embedding model name. See the table above |
input | string / array | Yes | A single string, or an array of strings for batch requests |
dimensions | integer | No | Output dimensions. Only text-embedding-3 models support it. Fewer dimensions save storage at a small cost in accuracy |
encoding_format | string | No | float (default) or base64. base64 is a Base64-encoded little-endian float32 array and is smaller to transfer |
| Field | Description |
|---|---|
data[].index | Position in input |
data[].embedding | The vector |
usage.prompt_tokens | Input tokens. Billing is based on this value |