Relevance Bench - nDCG@10 Kendall tau-b: leaderboard
Relevance Bench - nDCG@10 Kendall tau-b. TREC-style relevance judgments; 22327 available judgments, source reports partial coverage of22256,22324,22320 for three services. OpenAI Decisions is a separate scoring route; correlation and agreement, not answer accuracy. Metric: nDCG@10 Kendall tau-b (−1 to 1). Primary published table; source configurations remain distinct.
Metric: nDCG@10 Kendall tau-b (−1 to 1). Source: iwhalen.com. Saturation forecast: Around January 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Luna | 0.93 | #52 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=relevance-bench-ndcg-10-kendall-tau-b · How It Works · Data refreshed daily, snapshot 2026-10-16.