Question 1

What is Text Embedding V4?

Accepted Answer

Text Embedding V4: Qwen3-based text embedding model for retrieval and knowledge bases, served at /v1/embeddings.

Question 2

How much does Text Embedding V4 cost?

Accepted Answer

From ¥0.5 per million input tokens and ¥0 per million output tokens, at the provider's price. Credits add a platform fee when you top up; with your own provider keys there is no fee.

Question 3

Which providers serve Text Embedding V4?

Accepted Answer

Alibaba Qwen. DigitSpace sends each request to whichever of them is healthy and cheapest, and fails over automatically.

Question 4

How long a prompt can Text Embedding V4 take?

Accepted Answer

Up to 8,192 tokens of context.

Question 5

How do I call Text Embedding V4?

Accepted Answer

Use any OpenAI-compatible SDK with the base URL https://api.digitspace.cn/v1 and set model to "text-embedding-v4". To pin a provider, use "provider/text-embedding-v4".

Text Embedding V4

Providers

What users think

Questions

More from alibaba