New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
AI embedding models
Compare text embedding models for semantic search, retrieval-augmented generation (RAG) and NLP. Review embedding dimensions, context length, MTEB scores and API pricing.
LFM2.5-ColBERT-350M
🇺🇸 Liquid AI
LFM2.5-Embedding-350M
🇺🇸 Liquid AI
mxbai-rerank-v3-listwise
Mixedbread
jina-embeddings-v5-omni
Jina AI
Granite Embedding Multilingual R2
🇺🇸 IBM
granite-embedding-97m-multilingual-r2
🇺🇸 IBM
LaSER-Qwen3-0.6B
🇨🇳 Qwen
LaSER-Qwen3-4B
🇨🇳 Qwen
LaSER-Qwen3-8B
🇨🇳 Qwen
Harrier-OSS-v1-0.6B
🇺🇸 Microsoft
Harrier-OSS-v1-270M
🇺🇸 Microsoft
Harrier-OSS-v1-27B
🇺🇸 Microsoft
Wholembed v3
Mixedbread
Gemini Embedding 2
🇺🇸 Google DeepMind
Llama-NV-Embed-Reasoning-3B
🇺🇸 NVIDIA
LateOn-Code
🇫🇷 LightOn
LateOn-Code-edge
🇫🇷 LightOn
Ops-Colqwen3-4B
OpenSearch-AI
jina-embeddings-v5-text-nano
Jina AI
jina-embeddings-v5-text-small
Jina AI
voyage-4
🇺🇸 Voyage AI
voyage-4-large
🇺🇸 Voyage AI
voyage-4-lite
🇺🇸 Voyage AI
voyage-4-nano
🇺🇸 Voyage AI
voyage-multimodal-3.5
🇺🇸 Voyage AI
Qwen3-VL-Embedding-2B
🇨🇳 Qwen
Qwen3-VL-Embedding-8B
🇨🇳 Qwen
Qwen3-VL-Reranker-2B
🇨🇳 Qwen
Qwen3-VL-Reranker-8B
🇨🇳 Qwen
Amazon Nova Multimodal Embeddings
🇺🇸 Amazon
Seed1.5-Embedding
🇨🇳 ByteDance
Embed 4
🇨🇦 Cohere
About AI embedding models
Embedding models convert text into dense vectors that capture meaning, powering semantic search, retrieval-augmented generation (RAG), clustering and classification. The best embedding model balances retrieval quality against dimensionality, context length and cost. MTEB is the standard benchmark for ranking embedding quality across retrieval and similarity tasks, but practical choice also depends on vector dimension (which drives storage and search cost), maximum input length and whether you can self-host. Open-source embedding models let you run retrieval entirely on your own infrastructure, while hosted APIs simplify scaling. Browse the list below to compare embedding models by provider and availability, see rankings on our benchmarks page, and use compare to weigh two models for your RAG pipeline.
Frequently asked questions
How do I choose the right model?
Weigh raw capability against practical constraints like context window, latency, licensing and price. Use the benchmarks page to compare rankings and the compare tool to evaluate two candidates side by side.
Where can I see benchmark scores?
Visit the benchmarks page to compare these models on standardised tests, then use the compare tool for a detailed side-by-side of any two models.