EuroEval Albanian NLU - MultiWikiQA SQ: leaderboard
The "MultiWikiQA SQ" subtask of EuroEval Albanian NLU. Albanian-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Character F1 (%). Source: euroeval.com. Saturation forecast: Around January 2027. 200 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Mistral Small 3.2 | 70.5 | #605 |
| 2 | Ministral-3-14B-Reasoning-2512 | 70.23 | #540 |
| 3 | Qwen 3 14B (Non-reasoning) | 68.4 | #537 (Qwen 3 14B) |
| 4 | Qwen 3 32B (Non-reasoning) | 67.72 | #443 (Qwen 3 32B) |
| 5 | Voxtral-Small-24B-2507 | 67.29 | #581 |
| 6 | Llama 3.1 70B | 67.18 | #592 |
| 7 | Qwen3.5-4B-Base | 67.03 | #798 |
| 8 | gemma-3-12B-pt | 66.71 | #686 |
| 9 | gemma-3-27B-pt | 66.54 | #502 |
| 10 | Mistral Small 3.1 | 65.7 | #610 |
| 11 | Ministral-3-14B-Instruct-2512 | 65.11 | #615 |
| 12 | Qwen 3 8B (Non-reasoning) | 63.62 | #674 (Qwen 3 8B) |
| 13 | Llama-3.1-Tulu-3-8B-SFT | 62.93 | #985 |
| 14 | Ministral-3-8B-Instruct-2512 | 62.31 | #661 |
| 15 | Llama 3.1 8B | 61.84 | #1151 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-albanian-nlu-multiwikiqa-sq · How It Works · Data refreshed daily, snapshot 2026-10-20.