# qwen/qwen3-embedding-4b-batch

Calls to the model qwen/qwen3-embedding-4b-batch on DeepInfra stopped working on July 6, 2026.

DeepInfra retired model qwen/qwen3-embedding-4b-batch on July 6, 2026. Calls specifying this model fail since that date. The change took effect across all requests targeting the model on api.deepinfra.com.

## What to use instead

Move to Qwen/Qwen3-Embedding-4B, as recommended by DeepInfra. Switching requires more than changing the model identifier, because embeddings from another model are not comparable and stored vectors must all be computed again with the new model.

Embeddings from another model are not comparable with this one's: vectors the code stores must all be computed again with the new model.

- API: DeepInfra (api.deepinfra.com)
- Model: qwen/qwen3-embedding-4b-batch
- Kind: Model retirement
- Takes effect: July 6, 2026
- Calls after that day: fail
- Calls it hits: `model qwen/qwen3-embedding-4b-batch`
- Source: [Model list](https://api.deepinfra.com/models/list)
- Verified: November 6, 2026, against DeepInfra's own pages
- Page: https://www.okapi.bot/models/deepinfra/qwen/qwen3-embedding-4b-batch/

Okapi finds the calls this change hits in your repo and opens the pull request with the fix: https://github.com/apps/okapi-agent