Math Word Problem Generation on Standards-aligned Educational MWPs
9.5PPLEDUMATH 12B
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| EDUMATH 12BAccess Type=Public, Model Size=12B2025.10 | 9.5 | 74.5 | 73.8 | 54.9 | 166.7 | 0.389 | 0.381 | 85.9 | |
| Gemma 3 12B ITAccess Type=Public, Model Size=12B, Instruction Tuned=true2025.10 | 11.3 | 77.2 | 74 | 84.7 | 240.8 | 0.372 | 0.376 | 63.9 | |
| EDUMATH 30BAccess Type=Public, Model Size=30B2025.10 | 12 | 76.1 | 73.8 | 60.4 | 163.5 | 0.38 | 0.379 | 94.6 | |
| Gemma 3 27B ITAccess Type=Public, Model Size=27B, Instruction Tuned=true2025.10 | 12.2 | 77.1 | 74.5 | 76.1 | 215.1 | 0.371 | 0.378 | 75.4 | |
| Qwen 3 30B ITAccess Type=Public, Model Size=30B, Instruction Tuned=true2025.10 | 12.3 | 76.9 | 74 | 68.1 | 202.9 | 0.383 | 0.383 | 87.3 | |
| Qwen 3 235B ITAccess Type=Public, Model Size=235B, Instruction Tuned=true2025.10 | 12.5 | 76.1 | 74.1 | 66.5 | 186.5 | 0.392 | 0.378 | 89 | |
| GPT-4o-miniAccess Type=API2025.10 | 15.7 | 75 | 74.1 | 61.8 | 150.6 | 0.383 | 0.399 | 92 | |
| GPT-4oAccess Type=API2025.10 | 16.1 | 75.2 | 74.2 | 63.5 | 152.3 | 0.378 | 0.399 | 92.8 | |
| GPT-4.0Access Type=API2025.10 | 16.3 | 75.3 | 74.3 | 64.5 | 150.2 | 0.385 | 0.4 | 92.8 |