EuroEval Croatian NLU - WikiANN HR: leaderboard
The "WikiANN HR" subtask of EuroEval Croatian NLU. Croatian-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Micro F1 excluding MISC (%). Source: euroeval.com. Saturation forecast: Around January 2027. 211 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 27B FP8 | 70.4 | #164 |
| 2 | Claude Sonnet 4.5 (Thinking) | 69.26 | #142 (Claude Sonnet 4.5) |
| 3 | Gemini 3 Pro (Preview) | 68.98 | #66 |
| 4 | Gemma 4 31B (IT) | 68.87 | #176 |
| 5 | Mistral Small 3.2 | 68.75 | #605 |
| 6 | Voxtral-Small-24B-2507 | 67.95 | #581 |
| 7 | Qwen 3 4B 2507 (Thinking) | 67.93 | #538 (Qwen 3 4B 2507) |
| 8 | GPT-5.4 Mini (2026-03-17) (High) | 67.85 | #169 (GPT-5.4 Mini (2026-03-17)) |
| 9 | Mistral Small 3.1 | 67.69 | #610 |
| 10 | OLMo 3.1 32B (Thinking) | 67.54 | #632 (OLMo 3.1 32B) |
| 11 | Llama 3.1 70B | 67.53 | #592 |
| 12 | Qwen 3.5 35B A3B FP8 | 67.48 | #362 |
| 13 | MamayLM-Gemma-3-12B-IT-v1.0 | 67.14 | #584 |
| 14 | Qwen 3 14B (Non-reasoning) | 67.03 | #537 (Qwen 3 14B) |
| 15 | Gemma 4 26B A4B (IT) | 66.93 | #297 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-croatian-nlu-wikiann-hr · How It Works · Data refreshed daily, snapshot 2026-10-20.