EuroEval Croatian NLU - Multi Wiki QA HR: leaderboard

Metric: Reading comprehension Score (%). Source: euroeval.com. 211 models tracked.

Top models

#ModelScore
1Ministral-3-14B-Reasoning-251273.9
2Olmo-3.1-32B-Instruct-SFT70.12
3Ministral 3 14B69.62
4SOLAR-10.7B-v1.068.75
5Mistral Small 3.268.62
6Ministral-3-14B-Instruct-251267.67
7Gemini 3 Pro (Preview)67.64
8GPT-5.267.02
9Llama 3.1 70B66.7
10Ministral-3-3B-Reasoning-251266.43
11GPT-5.4 Mini (Medium)66.25
12Gemini 2.5 Flash (Thinking)65.65
13gemma-4-E2B-it65.47
14GPT-5 Nano65.17
15Olmo-3.1-32B-Instruct-DPO65.08

Interactive version: theaggregate.ai/benchmark?slug=euroeval-croatian-nlu-multi-wiki-qa-hr · How It Works · Data refreshed daily, snapshot 2026-09-05.