Zerank 2
State-of-the-art cross-encoder trained with novel zELO pipeline for calibrated relevance scores. First reranker with native instruction-following for domain-specific optimization, multilingual parity across 100+ languages, normalized scores with confidence statistics, and robust SQL-style aggregation query handling. If you want to compare the best rerankers for your data, try Agentset.
Model Information
- Provider
- ZeroEntropy
- License
- cc-by-nc-4.0
- Price per 1M tokens
- $0.025
- Release Date
- 2025-11-18
- Model Name
- zerank-2
- Total Evaluations
- 3300
Performance Record
Rerankers Are Just One Piece of RAG
Agentset gives you a managed RAG pipeline with the top-ranked models and best practices baked in. No infrastructure to maintain, no reranking to configure.
Trusted by teams building production RAG applications
Performance Overview
ELO ratings by dataset
Zerank 2's ELO performance varies across different benchmark datasets, showing its strengths in specific domains.
Zerank 2 - ELO by Dataset
Detailed Metrics
Dataset breakdown
Performance metrics across different benchmark datasets, including accuracy and latency percentiles.
FiQa
Accuracy Metrics
- nDCG@5
- 0.108
- nDCG@10
- 0.119
- Recall@5
- 0.098
- Recall@10
- 0.130
Latency Distribution
- Mean
- 251ms
- P50 (Median)
- 254ms
- P90
- 270ms
arguana
Accuracy Metrics
- nDCG@5
- 0.283
- nDCG@10
- 0.355
- Recall@5
- 0.540
- Recall@10
- 0.760
Latency Distribution
- Mean
- 280ms
- P50 (Median)
- 278ms
- P90
- 312ms
MSMARCO
Accuracy Metrics
- nDCG@5
- 0.000
- nDCG@10
- 0.000
- Recall@5
- 0.000
- Recall@10
- 0.000
Latency Distribution
- Mean
- 233ms
- P50 (Median)
- 228ms
- P90
- 251ms
PG
Accuracy Metrics
- nDCG@5
- 0.000
- nDCG@10
- 0.000
- Recall@5
- 0.000
- Recall@10
- 0.000
Latency Distribution
- Mean
- 293ms
- P50 (Median)
- 281ms
- P90
- 328ms
DBPedia
Accuracy Metrics
- nDCG@5
- 0.000
- nDCG@10
- 0.000
- Recall@5
- 0.000
- Recall@10
- 0.000
Latency Distribution
- Mean
- 247ms
- P50 (Median)
- 242ms
- P90
- 274ms
business reports
Accuracy Metrics
- nDCG@5
- 0.000
- nDCG@10
- 0.000
- Recall@5
- 0.000
- Recall@10
- 0.000
Latency Distribution
- Mean
- 288ms
- P50 (Median)
- 269ms
- P90
- 367ms
Build RAG in Minutes, Not Months
Agentset gives you a complete RAG API with top-ranked rerankers and embedding models built in. Upload your data, call the API, and get accurate results from day one.
import { Agentset } from "agentset";
const agentset = new Agentset();
const ns = agentset.namespace("ns_1234");
const results = await ns.search(
"What is multi-head attention?"
);
for (const result of results) {
console.log(result.text);
}Compare Models
See how it stacks up
Compare Zerank 2 with other top rerankers to understand the differences in performance, accuracy, and latency.