RAG indexing cost calculator
What document search for an AI assistant (RAG) costs in embeddings and reranking: your documents embedded once, the changed share every month, the queries, and the passages reranked - at the cheapest seller of every model. How to use it
Documents: 6 000 000 tokens. Every month: 1 500 000 tokens to embed and 258 000 000 tokens to rerank.
Embedding models
Rerank models
| Model | Provider | Per 1M tokens | A month |
|---|---|---|---|
| Qwen3 Reranker 0.6B | deepinfra | $0.01 | $2.58 |
| BGE Reranker V2 M3 | novita | $0.01 | $2.58 |
| Llama Nemotron Rerank VL 1B v2 | deepinfra | $0.01 | $2.58 |
| Qwen3-VL-Reranker-2B | moark | $0.01349 | $3.48 |
| Rerank 5 Fast | aihubmix | $0.02 | $5.16 |
| Qwen3 Reranker 4B | moark | $0.02098 | $5.41 |
| Qwen3 Reranker 8B | moark | $0.04197 | $10.83 |
| Cohere Rerank 5 Pro | aihubmix | $0.05 | $12.90 |
| jina-reranker-m0 | aihubmix | $0.05 | $12.90 |
| jina-reranker-v3 | aihubmix | $0.05 | $12.90 |
| jina-reranker-v3.5 | aihubmix | $0.05 | $12.90 |
| Nemotron-3.5-Rerank-8B-BF16 | deepinfra | $0.05 | $12.90 |
| rerank-2.5-lite | 302ai | $0.05 | $12.90 |
| Qwen3-VL-Reranker-8B | moark | $0.05246 | $13.53 |
| Rerank 4 Fast | aihubmix | $0.068 | $17.54 |
| Rerank 4 Pro | aihubmix | $0.068 | $17.54 |
| bce-reranker-base | aihubmix | $0.068 | $17.54 |
| Qwen3 Rerank | 302ai | $0.07 | $18.06 |
| Qwen: Qwen3 VL Rerank | zenmux | $0.10 | $25.80 |
| rerank-2.5 | 302ai | $0.10 | $25.80 |
| gte-rerank-v2 | aihubmix | $0.11 | $28.38 |
Counted: embedding the documents once, re-embedding the changed share and every query each month, reranking the passages of every query. Not counted: the vector database, storage, and the answer model that writes from the passages - price it in the cost calculator or the coding agent calculator.