General Language Understanding and Reasoning on General domain benchmarks (test)
93.3DROP ScoreAM-Thinking (math)
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| AM-Thinking (math)Backbone=Qwen3-8B-Base, Size=558k2025.12 | 93.3 | 0.013 | 38.7 | 74.7 | 56.9 | 65.9 | |
| ODA-Mixture-100kBackbone=Qwen3-8B-Base, Size=101k2025.12 | 91.6 | 0.024 | 44.2 | 73.3 | 35.1 | 61.1 | |
| AM-Thinking (code)Backbone=Qwen3-8B-Base, Size=324k2025.12 | 88.2 | 0.019 | 50.1 | 70.4 | 50.8 | 64.9 | |
| ODA-Mixture-100kBackbone=Qwen2.5-7B-Base, Size=101k2025.12 | 87.3 | 0.053 | 39.1 | 61.2 | 39.6 | 56.8 | |
| MiroMind-M1-SFTBackbone=Qwen3-8B-Base, Size=719k2025.12 | 85 | 0.008 | 43.5 | 74.2 | 55.1 | 64.5 | |
| Light-R1-SFTBackbone=Qwen3-8B-Base, Size=79k2025.12 | 83.4 | 0.078 | 46.8 | 71.6 | 57.7 | 64.9 | |
| Raiden-DeepSeek-R1Backbone=Qwen3-8B-Base, Size=63k2025.12 | 82.3 | 0.143 | 63.2 | 71.9 | 53.3 | 67.7 | |
| TextbookReasoningBackbone=Qwen2.5-7B-Base, Size=652k2025.12 | 79.7 | 0.018 | 49.2 | 67.7 | 54.8 | 62.9 | |
| Magpie-Reasoning-V1Backbone=Qwen2.5-7B-Base, Size=150k2025.12 | 78.4 | 0.065 | 55.5 | 63.3 | 47.4 | 61.1 | |
| ODA-Mixture-500kBackbone=Qwen3-8B-Base, Size=506k2025.12 | 77.4 | 0.025 | 71.2 | 75.3 | 60.8 | 71.2 | |
| tulu-3-sft-personas-algebraBackbone=Qwen2.5-7B-Base, Size=20k2025.12 | 76.9 | 0.482 | 48.9 | 65.6 | 52.9 | 61.1 | |
| MegaScienceBackbone=Qwen2.5-7B-Base, Size=1.25M2025.12 | 75.9 | 0.008 | 48.1 | 65.3 | 57.7 | 61.8 | |
| ODA-Mixture-500kBackbone=Qwen2.5-7B-Base, Size=506k2025.12 | 75.2 | 0.024 | 57.7 | 66 | 54.8 | 63.4 | |
| math-gpt-4o-200kBackbone=Qwen2.5-7B-Base, Size=200k2025.12 | 74.8 | 0.055 | 52.4 | 68.8 | 53.8 | 62.5 | |
| Qwen3-8B-BaseBackbone=Qwen3-8B-Base2025.12 | 71.5 | — | 45.9 | 61.1 | 56.2 | 58.7 | |
| Qwen2.5-7B-BaseBackbone=Qwen2.5-7B-Base2025.12 | 68.3 | — | 35.5 | 57.7 | 44.2 | 51.4 |