Mathematical Reasoning on GlobalMGSM 35 languages
62.5Mean AccuracyQwen3.5-4B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3.5-4BParameters=4B2026.03 | 62.5 | 26.7 | |
| Qwen3-4BParameters=4B2026.03 | 60.9 | 33.6 | |
| Gemma3-4BParameters=4B2026.03 | 55.4 | 26.5 | |
| Tiny Aya Global2026.03 | 52.8 | 11.7 | |
| Ministral-3-3BParameters=3B2026.03 | 49.6 | 34 | |
| SmolLM3-3BParameters=3B2026.03 | 22.6 | 27 |