Qwen3 Embedding
Awaiting activationAn open-weight multilingual model for search over text and technical information.
What it does
Qwen3 Embedding 0.6B is the compact member of Qwen’s open-weight embedding family. Its official model card describes support for more than 100 languages, including programming languages, and instruction-aware multilingual retrieval.
The model has a 32,000-token source context and output dimensions up to 1,024. In embeddings.app this entry is reserved for an explicitly configured Databricks serving deployment. Open weights describe how the model is distributed; they do not mean this hosted service runs for free or is already active.
Where it helps
Multilingual knowledge
Find related information across a diverse text collection.
Technical search
Evaluate retrieval over technical documentation and source-related text.
Existing deployments
Use a deliberately configured Qwen3 serving endpoint through the shared API.
When to choose it
Consider it if you already operate the matching Databricks deployment and want a consistent embedding interface. For an immediately usable model on this site, choose an entry marked Available.
Awaiting a configured deployment and verified account pricing. This exact entry is the 0.6B model, not the 4B or 8B variants; their results and dimensions should not be treated as interchangeable.
Official sources
Paraphrased from the provider’s documentation and original model information. Selection advice is our editorial guidance.
Qwen · Qwen3 Embedding 0.6B original model card ↗Reviewed October 3, 2026