EuroEval Albanian NLU - WikiANN SQ — leaderboard

Metric: Named entity recognition Score (%). Source: euroeval.com. 200 models tracked.

Top models

#ModelScore
1GPT-5.4 Mini (Medium)81.85
2GPT-5.4 Mini (High)81.55
3Claude Sonnet 4.5 (Thinking)81.3
4Ministral 3 14B81.01
5Gemini 3.1 Pro (Preview)80.88
6Gemma 4 31B (IT)80.13
7Gemini 3 Pro (Preview)79.6
8gemma-3-27B-pt79.15
9GLM-4.7 Flash79.08
10gemma-3-12B-pt78.97
11Mistral Small 3.278.7
12Ministral 3 8B78.68
13GPT-578.58
14Gemma 4 26B A4B (IT)78.33
15Grok 4.20 Beta (0309) (Reasoning)77.99

Interactive version: theaggregate.ai/benchmark?slug=euroeval-albanian-nlu-wikiann-sq · How the rankings work · Data refreshed daily, snapshot 2026-07-22.