Tool-augmented planning on MagicEval-Tool Dependency
98.5Name AccMagicAgent-30B-A3B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MagicAgent-30B-A3BScale Category=MagicAgent Series2026.02 | 98.5 | 78.5 | |
| MagicAgent-32BScale Category=MagicAgent Series2026.02 | 97.5 | 80 | |
| Kimi-K2-InstructScale Category=Ultra-Scale2026.02 | 91 | 58.3 | |
| Qwen3-MAXScale Category=Ultra-Scale2026.02 | 90 | 58.3 | |
| Qwen3-235B-A22B-InstructScale Category=Ultra-Scale2026.02 | 89.5 | 55.3 | |
| DeepSeek-V3.1-nothinkScale Category=Ultra-Scale2026.02 | 86.8 | 54.8 | |
| GLM-4.7Scale Category=Ultra-Scale2026.02 | 85.3 | 56.5 | |
| Llama3.3-70B-InstructScale Category=Large-Scale2026.02 | 71 | 41.8 | |
| Qwen3-32B-nothinkScale Category=Large-Scale2026.02 | 69.8 | 43.8 | |
| GPT-5.2Scale Category=Ultra-Scale2026.02 | 67.8 | 46.5 | |
| ERNIE-4.5-21B-A3B-PTScale Category=Large-Scale2026.02 | 61.5 | 32.8 | |
| Qwen3-30B-A3B-InstructScale Category=Large-Scale2026.02 | 49.3 | 30 | |
| GLM-4.7-FlashScale Category=Large-Scale2026.02 | 24.8 | 13 | |
| Olmo-3.1-32B-InstructScale Category=Large-Scale2026.02 | 21.3 | 9.5 |