EuroEval Serbian NLU - MMS SR: leaderboard
The "MMS SR" subtask of EuroEval Serbian NLU. Serbian-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Matthews correlation coefficient (%). Source: euroeval.com. Saturation forecast: Around April 2027. 203 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.3 70B Instruct | 52.43 | #534 |
| 2 | Gemma 3 27B (IT) | 51.83 | #523 |
| 3 | MamayLM-Gemma-3-12B-IT-v1.0 | 51.48 | #584 |
| 4 | Mistral Small 3.1 | 51.2 | #610 |
| 5 | Llama 3.1 70B | 51.11 | #592 |
| 6 | Gemma 3 12B (IT) | 50.97 | #664 |
| 7 | Qwen 3 30B A3B 2507 Instruct | 50.75 | #488 |
| 8 | Gemma 3n E4B (IT) | 50.75 | #749 |
| 9 | Mistral Small 3.2 | 50.4 | #605 |
| 10 | Qwen 2.5 72B Instruct | 50.39 | #454 |
| 11 | Gemini 3 Flash (Preview) (Thinking) | 50.38 | #82 (Gemini 3 Flash (Preview)) |
| 12 | lapa-v0.1.2-instruct | 50.24 | #672 |
| 13 | Qwen 3 32B (Non-reasoning) | 49.99 | #443 (Qwen 3 32B) |
| 14 | Apertus-70B-Instruct-2509 | 49.86 | #820 |
| 15 | OLMo-2-0325-32B-Instruct | 49.82 | #848 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-serbian-nlu-mms-sr · How It Works · Data refreshed daily, snapshot 2026-10-20.