Rerank cost calculator
Hosted rerankers do not bill the same way. Cohere charges per search — one query plus up to 100 documents, whatever their length. Voyage charges per token. So the cheaper vendor flips depending on how long your passages are and how many you rerank. Put your own numbers in.
Your workload
Monthly cost
| Option | Billing unit | Billable volume | Monthly cost |
|---|
Why the ranking changes
Per-search pricing is indifferent to passage length: reranking 50 chunks of 80 tokens costs exactly what 50 chunks of 800 tokens costs. Per-token pricing scales with everything you send. That gives a simple rule of thumb:
- Short passages, large top-k — per-token billing tends to win, because you are paying for very little text.
- Long passages — per-search billing tends to win, and its advantage grows with every extra token in the chunk.
- Top-k above 100 — Cohere starts a second billable search per query, so cost steps up rather than sliding.
Breakeven: —
What about self-hosting?
Self-hosting has no per-call price, so it does not belong in the table above — you trade a usage bill for a GPU bill plus operations. The comparison only becomes meaningful at your own volume: divide your monthly GPU cost by the number of queries above and compare that to the per-query figures in the table. Below a few hundred thousand queries a month an API is usually cheaper than a dedicated GPU; well above that, self-hosting a Qwen3-Reranker or bge-reranker-v2-m3 starts to pay off. Our self-hosting guide covers the serving side.
Note that a smaller model is not automatically a worse one — Jina Reranker v3 reaches 61.94 BEIR nDCG@10 at 0.6B, so the cheapest thing to serve may also be the most accurate.
Assumptions and sources
- Cohere Rerank 4 — $0.0025 per search (Pro) and $0.002 per search (Fast), where a search is one query plus up to 100 documents. More than 100 candidates bills as multiple searches.
- Voyage rerank-2.5 — $0.05 per 1M tokens, and $0.02 per 1M tokens for
rerank-2.5-lite. Billable tokens are the query plus every document you send. The first 200M tokens per account are free and the Batch API is discounted 33%; neither is applied above. - Token counts are approximate — every vendor tokenises differently, and this page multiplies your averages rather than tokenising real text.
- Rates verified August 2026. Check Cohere and Voyage for current pricing before committing to a budget.
Cost is only half the decision
Quality is the other half. Run a cross-encoder against your own passages in the browser — no key, no spend.
Open the live demo →