Skip to main content
Embedding

Embedding & reranking models

Choose a model for semantic search, RAG retrieval, cross-modal matching, and reranking.

Model selection

Choose a model based on your input data type and use case.
  • Text or code — Use text-embedding-v4. Best general-purpose text embedding model, supporting task instructions (instruct), sparse vectors, and other advanced features.
  • Multimodal content — In Singapore, use tongyi-embedding-vision-plus for cross-modal retrieval (text-to-image, image-to-image, text-to-video). For cost-sensitive scenarios, use tongyi-embedding-vision-flash. In China North 2 (Beijing), use qwen3-vl-embedding, which returns fused or independent vectors for text, images, and videos.
  • Large-scale data — Use qwen3.7-text-embedding or text-embedding-v4 with Batch API to significantly reduce costs.

How many dimensions?

Large-scale search where storage matters → 256 or 512. General use → 1024 (default, good balance). Maximum accuracy on benchmarks → 1536 or 2048.

Text-only data?

Use text-embedding-v4 instead — faster, cheaper, more dimension options. Multimodal embedding is for cross-modal retrieval (text-to-image, text-to-video).

Accuracy vs speed

Best accuracy → tongyi-embedding-vision-plus (dimensions up to 1152). Budget or latency-sensitive → tongyi-embedding-vision-flash (up to 768).

Reranking

Improve RAG precision → add qwen3-rerank after your embedding search. Re-scores top-N results with cross-attention for better ranking quality. Limits: 500 documents per request, 4,000 tokens per item, 120,000 tokens per request. To rerank images or videos, use qwen3-vl-rerank (China North 2 (Beijing) only). It reranks mixed text, image, and video candidates and is called through the DashScope SDK or HTTP endpoint, not the OpenAI-compatible endpoint. Limits: 100 text documents, 40 images, or 4 videos per request; 8,000 tokens per document; 120,000 tokens per request.

Migrate from closed-source models

Replacing OpenAI, Cohere, or Voyage embeddings? Use these QwenCloud equivalents.
TaskClosed-source examplesQwenCloud recommendation
Text embeddingOpenAI text-embedding-3-large, Voyage-3-large, Cohere embed-v4text-embedding-v4
Multimodal embeddingCohere embed-v4, Voyage multimodaltongyi-embedding-vision-plus
RerankingCohere Rerank 3.5, Voyage rerank-2.5qwen3-rerank (Singapore, China North 2 (Beijing)), qwen3-vl-rerank (China North 2 (Beijing))

All models

In the console, the model list, search results, versions, prices, and rate limits follow the current region. Models differ by region, and some are available only in specific regions, so you can only compare models within the same region. Before you choose a model, check the region's tab in Model prices. See Regions and endpoints.
The model pages in the console, including model details and model comparison, show the models, prices, and rate limits of the region selected in the upper-right corner. A model offered in one region may not be listed in another. If you switch regions while a model detail page or the comparison list is open, a model that is not available in the new region is flagged, and the page shows the other regions where the model is available. The comparison page stays open even if only one model remains.
ModelRegionsUse this when...DimensionsMax tokens
text-embedding-v4Singapore, China North 2 (Beijing), China (Hong Kong)Text search, RAG, clustering64, 128, 256, 512, 768, 1024 (default), 1536, 20488,192
text-embedding-v3SingaporeExisting v3 index migration512, 768, 1024 (default)8,192
tongyi-embedding-vision-plusSingaporeCross-modal search, best accuracy1152, 1024, 512, 256, 128, 641,024
tongyi-embedding-vision-flashSingaporeCross-modal search, budget768, 512, 256, 128, 641,024
qwen3-vl-embeddingChina North 2 (Beijing)Multimodal search with fused or independent vectors2560 (default), 2048, 1536, 1024, 768, 512, 25632,000
qwen3-rerankSingapore, China North 2 (Beijing)Re-rank search results—4,000/item
qwen3-vl-rerankChina North 2 (Beijing)Re-rank text, image, and video results—8,000/item

Learn more