Embedding models for semantic search

For non-Latin text the tokenizer matters as much as the vector quality, so prefer a multilingual model and test it on your own documents. Embeddings are billed on input tokens only, which makes the one-off indexing pass cheap; the recurring cost sits in the answering step, not here.

0 models in this category · prices in toman per 1M tokens

Model	Capabilities	Context	Input / 1M	Output / 1M	≈ 1,000 words
No model matches this category right now. The catalogue syncs with the provider every hour, so this list changes.

Other categories

Models that hold up in a real codebase Cheap models for high-volume work Free models, shared capacity Models that read images Reasoning models, and what they cost Long-context models for documents and codebases Models that call your tools reliably Models that return images Batch variants (cheaper, slower)

See this category ranked by real usage →