Model pricing

Compare embedding API prices.

455 language models, 37 video models, 55 image models, 100 embedding and rerank models and 2,990 offers from the AI labs, clouds and hosts: what each costs, who sells it cheapest and how it ranks on LMArena. Enter your usage to see a monthly estimate.

As of Oct 19, 2026, the cheapest embedding API is pplx-embed-v1-0.6b at $0.004 per million tokens from Perplexity; OpenAI's text-embedding-3-small costs $0.02.

Estimate your monthly cost

Est. per month = tokens × price per million tokens, at the cheapest seller. Embedding models charge for input only. A million tokens is roughly 750,000 English words.

77 embedding models · prices per million tokens from the cheapest seller; open a row for every seller

Rerank models (23)

A reranker scores documents against a query to reorder search results. Sellers price it either per query (one query over a set of documents) or per token, so prices are compared only within the same unit.

About this data

Prices come from LiteLLM's model cost map, with more sellers from models.dev (both MIT licence), checked Oct 19, 2026. Providers change prices without notice; confirm on their pricing page before you rely on a number. Arena scores: LMArena text leaderboard of 2026-10-15 (CC BY 4.0). Provider logos: Lobe Icons.