Rerank retrieved chunks for RAG
After vector search, need a dedicated reranker so the agent sees the right 5–20 chunks — not just 'better embeddings'.
At a glance — Oct 2026
Updated 2026-10-27 · forward this page: /jtbds/rag-rerank#briefScores as of 2026-09-14 (43 days old; Depot marks scores stale after 30 days). The pick is still Depot's current editorial pick; it has not been re-reviewed since. Prices below carry their own dates.
- Depot Pick
- Voyage AI · 8.40/10 for this job (curated estimate, high confidence)
Capability of rerank quality + pricePerf (lite SKUs, free tokens). Distinct from embeddings (first-stage vectors).
- Runner-up
- Cohere · 7.95/10
Scores higher than the pick on: Privacy / data retention, Maturity / track record.
- When not to use the pick
- Already on Cohere embed+rerank in one bill — stay; want EU all-in-one embed/rerank/Reader — Jina; tiny corpus where embeddings alone suffice — skip rerank.
- Lock-in / migration
- The pick has lower stored lock-in risk than runner-up Cohere (4.2 vs 4.5). Self-host fallback listed for this job: BGE-reranker-v2-m3, Qwen3-Reranker-0.6B, FlashRank.
- Sourced price (usd_per_million_tokens)
Pick's lowest sourced SKU: rerank-3-lite $0.02 (usd_per_million_tokens) — source, as of 2026-10-03, re-checked 2026-10-27
Cheapest matching SKU: Voyage AI rerank-3-lite $0.02 — source, as of 2026-10-03, re-checked 2026-10-27. The pick is the cheapest.
Price history: first seen 2026-10-03 · last changed no change recorded · history JSON
- Free option
- Failure modes to watch
- Stale retrieval: Store the source URL and the time it was fetched. If a newer canonical page exists, or two stored sources disagree, do not answer from the older one.
- Runaway spend: Set a budget and a tool-call cap per task. Alert when spend jumps. Stop the run at the cap instead of summarizing a partial trace as success.
- Context overflow: If the input exceeds the window, say so and retrieve a narrower slice. Do not answer from a truncated head as if it were the whole source.
- Eval contamination: Keep gold labels out of the agent prompt. Store the harness version beside the score. Do not compare a number across a benchmark revision without saying the revision changed.
- Lost in the middle: Ask the agent to name the passage it used. Spot-check questions whose only support sits in the middle of the context, not only at the edges.
- Policy conflict: If cited sources contradict each other, return the conflict and the URLs. Do not pick a winner silently.
Cite: Agent Provider Depot, Depot Pick for rag-rerank: voyage, scores as of 2026-09-14. Agents: GET /api/decide?jtbd=rag-rerank
Free Rerank retrieved chunks for RAG options · All sourced prices, cheapest first · for rag agents
Voyage AI
Capability of rerank quality + pricePerf (lite SKUs, free tokens). Distinct from embeddings (first-stage vectors).
Pricing (directional): As of 2026-09-14 (docs.voyageai.com/docs/pricing): voyage-4-lite $0.02/1M, voyage-4 $0.06, voyage-4-large / voyage-context-4 / voyage-code-4 $0.12 (200M free tokens on those families). rerank-3 $0.05/1M and rerank-3-lite $0.02/1M (200M free…
Lock-in risk: higher = more lock-in (worse). Overall already inverts this dimension.
Depot pick for RAG rerank. Keep Cohere as general embeddings default (enterprise + multilingual maturity); switch to Voyage when retrieval nDCG dominates and you want the cheap lite SKU + rerank-3 in one vendor.
Quote for Rerank retrieved chunks for RAG
Depot Pick stays Voyage AI. Prices below are sourced SKU rows only (as of 2026-10-27). Input and output are not blended. GET /api/quote?jtbd=rag-rerank
Only 1 candidate has a matching Rerank $/1M; nothing was compared, so pickIsPriceWinner is null.
- rerank-3 usd_per_million_tokens $0.05 source as of 2026-10-03
- rerank-3-lite usd_per_million_tokens $0.02 source as of 2026-10-03
- voyage-4 usd_per_million_tokens $0.06 source as of 2026-10-03
- voyage-4-large usd_per_million_tokens $0.12 source as of 2026-10-03
- voyage-4-lite usd_per_million_tokens $0.02 source as of 2026-10-03
- voyage-code-4 usd_per_million_tokens $0.12 source as of 2026-10-03
- voyage-context-4 usd_per_million_tokens $0.12 source as of 2026-10-03
0 of 3 candidates have no sourced price row. 2 are excluded from the Rerank $/1M comparison (rerank). min sourced rerank SKU among candidates; per million tokens; not a chat model; not quality-adjusted
Runners-up
When NOT to use the top pick
Already on Cohere embed+rerank in one bill — stay; want EU all-in-one embed/rerank/Reader — Jina; tiny corpus where embeddings alone suffice — skip rerank.
Full ranking (score table)
Ranked with this JTBD's weights. Scores are curated estimates unless marked otherwise. Agents: prefer /api/decide then this table for tie-breaks.
Score methodology for this job
Capability of rerank quality + pricePerf (lite SKUs, free tokens). Distinct from embeddings (first-stage vectors).
capabilityFit: 25%pricePerformance: 18%latencyReliability: 15%developerDx: 15%privacyPosture: 10%lockInRisk: 7%maturity: 10%