Qwen 2.5 72B Instruct (Q4_K_M): benchmark results
Provider: Alibaba.
Aggregate Rating 1625 ± 52, rank #736 of 10870 rated models, from 6 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Polymorphic Sybil RAG | 18.9 | Gold retention (%; share of the 2,982 released HotpotQA and | 100 |
| Polymorphic Sybil RAG (Attack) | 34.9 | Gold retention (%; share of the 2,982 released HotpotQA and | 100 |
| Polymorphic Sybil RAG (Clean) | 46.7 | Gold retention (%; share of the 2,982 released HotpotQA and | 100 |
| Wolfram LLM Benchmarking Project | 49 | Correct Functionality (%) | 60.1 |
| Polymorphic Sybil RAG (Attack) - Hijack Rate | 17.6 | Hijack rate (%; share of the 2,982 released HotpotQA and Nat | 25 |
| Polymorphic Sybil RAG - Hijack Rate | 21.2 | Hijack rate (%; share of the 2,982 released HotpotQA and Nat | 25 |
Interactive version: theaggregate.ai/model?slug=qwen-2-5-72b-instruct-q4-k-m · How It Works · Data refreshed daily, snapshot 2026-11-12.