EuroEval Serbian NLU - MultiWikiQA SR: leaderboard
Metric: Reading comprehension Score (%). Source: euroeval.com. 205 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Ministral-3-14B-Reasoning-2512 | 68.07 |
| 2 | Ministral 3 14B | 67.72 |
| 3 | Mistral Small 3.2 | 67.16 |
| 4 | gemma-3-27B-pt | 66.09 |
| 5 | Llama 3.1 70B | 65.97 |
| 6 | GPT-5 Mini | 65.59 |
| 7 | GPT-5.4 Mini (Medium) | 65.23 |
| 8 | GPT-5.2 | 65 |
| 9 | Gemini 3 Pro (Preview) | 64.13 |
| 10 | SOLAR-10.7B-v1.0 | 64.1 |
| 11 | Grok 4.1 Fast (Reasoning) | 63.97 |
| 12 | Ministral 3 3B | 63.89 |
| 13 | Claude Sonnet 4.5 (Non-reasoning) | 63.43 |
| 14 | Gemini 2.5 Flash (Thinking) | 63.24 |
| 15 | GPT-5.4 Mini (High) | 63.23 |
Interactive version: theaggregate.ai/benchmark?slug=euroeval-serbian-nlu-multiwikiqa-sr · How It Works · Data refreshed daily, snapshot 2026-09-05.