| instruction-following-on-ifeval | Qwen3-14B + NGM | #13 | Prompt-level strict-accuracy: 83.92 |
| code-generation-on-livecodebench-v5 | Qwen3-14B + NGM | #19 | Pass@1: 29.94 |
| vision-language-understanding-on-mmbench-v1-1 | Qwen3-VL-2B + NGM | #23 | Accuracy: 76.63 |
| code-generation-on-humaneval | Qwen3-14B + NGM | #34 | Pass@1: 88.41 |
| optical-character-recognition-ocr-on-ocrbench | Qwen3-VL-2B + NGM | #43 | Score: 829 |
| arithmetic-reasoning-on-gsm8k | Qwen3-14B + NGM | #57 | Accuracy: 91.74 |
| mathematical-reasoning-on-math-500 | Qwen3-14B + NGM | #64 | Accuracy: 87.0 |
| vision-language-understanding-on-mmstar | Qwen3-VL-2B + NGM | #87 | Accuracy: 56.2 |
| multiple-choice-question-answering-mcqa-on-gpqa-diamond | Qwen3-14B + NGM | #111 | Accuracy: 52.02 |