EVALITA - word-in-context — leaderboard
Metric: CPS. Source: huggingface.co. 48 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 VL 32B Instruct | 77.56 |
| 2 | Mistral Large 2 (Nov) Instruct (2411) | 75.55 |
| 3 | calme-3.2-instruct-78B | 73.44 |
| 4 | Qwen 2.5 72B Instruct | 72.06 |
| 5 | Mistral Small 3 | 71.65 |
| 6 | Qwen 3 Next 80B A3B Instruct | 71.41 |
| 7 | DeepSeek R1 Distill Llama 70B | 69.26 |
| 8 | Phi-3-medium-4k-instruct | 68.92 |
| 9 | Gemma 2 27B (IT) | 68.56 |
| 10 | Llama 3.3 70B Instruct | 67.15 |
| 11 | Gemma 3 12B (IT) | 66.5 |
| 12 | Gemma 3 27B (IT) | 66.49 |
| 13 | Llama 3 8B 4bit UltraChat Ita | 66.22 |
| 14 | Phi-3.5-mini-instruct | 65.33 |
| 15 | granite-3.1-8B-instruct | 65.03 |
Interactive version: theaggregate.ai/benchmark?slug=evalita-word-in-context · How the rankings work · Data refreshed daily, snapshot 2026-07-22.