# Multitask Language Understanding on MMLU-ProX non-EU languages (test)

Source: [Multitask Language Understanding on MMLU-ProX non-EU languages (test)](<https://www.sota2.com/research/sota/multitask-language-understanding-on-mmlu-prox-non-eu-languages-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:30:51.359Z

Evaluation of model performance on the MMLU-ProX dataset, specifically focusing on non-EU languages within the test split, for Multitask Language Understanding, reporting Accuracy.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Multitask Language Understanding |
| Dataset (including any stated split) | MMLU-ProX non-EU languages (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 16 of 16 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Accuracy](<https://www.sota2.com/research/sota/multitask-language-understanding-on-mmlu-prox-non-eu-languages-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen-3-30B-A3B | 70.9 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Qwen-3-32B | 69 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Qwen-3-14B | 64.6 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Llama-3.3-70B | 63.4 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Mistral-3.2-24B | 62.5 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Gemma-3-27B | 58.8 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| OLMo-3.1-32B | 55.1 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Gemma-3-12B | 51.8 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| EuroLLM-22B (new) | 43.8 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| OLMo-3-7B | 40.5 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| EuroLLM-9B (new) | 36.5 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| EuroLLM-22B (old) | 36.4 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Apertus-70B | 35.2 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Llama-3.1-8B | 31.1 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| Apertus-8B | 28.6 | Not recorded | 698575d8c97a0b3e3b5d5e8c |
| EuroLLM-9B (old) | 27.9 | Not recorded | 698575d8c97a0b3e3b5d5e8c |

### Qwen-3-30B-A3B

Accuracy: 70.9. Result ID: 699647a15698665ec48b9910.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Release Type":"Open-weights","Model Category":"Non-European"}. This record does not assert independent reproduction.

### Qwen-3-32B

Accuracy: 69. Result ID: 699647a15698665ec48b990f.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Release Type":"Open-weights","Model Category":"Non-European"}. This record does not assert independent reproduction.

### Qwen-3-14B

Accuracy: 64.6. Result ID: 699647a15698665ec48b990e.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Release Type":"Open-weights","Model Category":"Non-European"}. This record does not assert independent reproduction.

### Llama-3.3-70B

Accuracy: 63.4. Result ID: 699647a15698665ec48b990b.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75d11e209131cea6a028.md>)

[Implementation](<https://github.com/meta-llama/llama-agentic-system>)

Recorded evaluation configuration: {"Release Type":"Open-weights","Model Category":"Non-European"}. This record does not assert independent reproduction.

### Mistral-3.2-24B

Accuracy: 62.5. Result ID: 699647a15698665ec48b9909.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70d61e209131cea4d909.md>)

[Implementation](<https://github.com/mistralai/mistral-src>)

Recorded evaluation configuration: {"Release Type":"Open-weights","Model Category":"European"}. This record does not assert independent reproduction.

### Gemma-3-27B

Accuracy: 58.8. Result ID: 699647a15698665ec48b990d.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa779f1e209131cea746cd.md>)

Recorded evaluation configuration: {"Release Type":"Open-weights","Model Category":"Non-European"}. This record does not assert independent reproduction.

### OLMo-3.1-32B

Accuracy: 55.1. Result ID: 699647a15698665ec48b9908.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/694238e845614693628503e0.md>)

Recorded evaluation configuration: {"Release Type":"Fully-open","Model Category":"Non-European"}. This record does not assert independent reproduction.

### Gemma-3-12B

Accuracy: 51.8. Result ID: 699647a15698665ec48b990c.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa779f1e209131cea746cd.md>)

Recorded evaluation configuration: {"Release Type":"Open-weights","Model Category":"Non-European"}. This record does not assert independent reproduction.

### EuroLLM-22B (new)

Accuracy: 43.8. Result ID: 699647a15698665ec48b9904.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

Recorded evaluation configuration: {"Release Type":"Fully-open","Model Category":"European"}. This record does not assert independent reproduction.

### OLMo-3-7B

Accuracy: 40.5. Result ID: 699647a15698665ec48b9907.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/694238e845614693628503e0.md>)

Recorded evaluation configuration: {"Release Type":"Fully-open","Model Category":"Non-European"}. This record does not assert independent reproduction.

### EuroLLM-9B (new)

Accuracy: 36.5. Result ID: 699647a15698665ec48b9902.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

Recorded evaluation configuration: {"Release Type":"Fully-open","Model Category":"European"}. This record does not assert independent reproduction.

### EuroLLM-22B (old)

Accuracy: 36.4. Result ID: 699647a15698665ec48b9903.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

Recorded evaluation configuration: {"Release Type":"Fully-open","Model Category":"European"}. This record does not assert independent reproduction.

### Apertus-70B

Accuracy: 35.2. Result ID: 699647a15698665ec48b9906.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b6c1e209131cea885c7.md>)

Recorded evaluation configuration: {"Release Type":"Fully-open","Model Category":"European"}. This record does not assert independent reproduction.

### Llama-3.1-8B

Accuracy: 31.1. Result ID: 699647a15698665ec48b990a.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75d11e209131cea6a028.md>)

[Implementation](<https://github.com/meta-llama/llama-agentic-system>)

Recorded evaluation configuration: {"Release Type":"Open-weights","Model Category":"Non-European"}. This record does not assert independent reproduction.

### Apertus-8B

Accuracy: 28.6. Result ID: 699647a15698665ec48b9905.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b6c1e209131cea885c7.md>)

Recorded evaluation configuration: {"Release Type":"Fully-open","Model Category":"European"}. This record does not assert independent reproduction.

### EuroLLM-9B (old)

Accuracy: 27.9. Result ID: 699647a15698665ec48b9901.

[Reporting paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698575d8c97a0b3e3b5d5e8c.md>)

Recorded evaluation configuration: {"Release Type":"Fully-open","Model Category":"European"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
