EVALITA - relation-extraction — leaderboard

Metric: CPS. Source: huggingface.co. 48 models tracked.

Top models

#ModelScore
1Mistral Large 2 (Nov) Instruct (2411)66.82
2Gemma 2 9B (IT)51.26
3Gemma 3n E4B (IT)49.65
4Gemma 3 27B (IT)48.48
5Qwen2.5-14B-Instruct-1M44.23
6LLaMAntino-3-ANITA-8B-Inst-DPO-ITA42.14
7Qwen 2.5 72B Instruct41.6
8Gemma 3 12B (IT)41.59
9Mistral Small 340.9
10granite-3.1-8B-instruct39.91
11Qwen 3 Next 80B A3B Instruct38.87
12Qwen 2.5 VL 32B Instruct38.41
13DeepSeek R1 Distill Llama 70B37.34
14Phi-437.15
15Phi-3-medium-4k-instruct36.95

Interactive version: theaggregate.ai/benchmark?slug=evalita-relation-extraction · How the rankings work · Data refreshed daily, snapshot 2026-07-22.