Agricultural Agent Planning and Execution on AgriAgent Non-Zero
87.2Task FulfillmentAgriAgent
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AgriAgentModel=Qwen3-235B2026.01 | 87.2 | — | 81.9 | 79.8 | |
| AgriAgentModel=DeepSeek-v3.12026.01 | 85.1 | — | 80.2 | 75.3 | |
| AgriAgentModel=Qwen2.5-14B2026.01 | 80.3 | 78.6 | 72.9 | 70.3 | |
| AgriAgentModel=Qwen3-32B2026.01 | 78.5 | 79.6 | 78.4 | 71.3 | |
| AgriAgent (w/o feedback)Model=Qwen3-32B2026.01 | 77.7 | 78.8 | 77 | 70.6 | |
| AgriAgent (w/o feedback)Model=Qwen2.5-14B2026.01 | 77.6 | 78 | 72.4 | 66.8 | |
| AgriAgentModel=Qwen2.5-7B2026.01 | 76.3 | 70.2 | 67.4 | 64.1 | |
| AgriAgent (w/o feedback)Model=AgriGPT-8B2026.01 | 75.1 | 77.6 | 72.9 | 68.8 | |
| AgriAgentModel=AgriGPT-8B2026.01 | 74.5 | 79 | 66.9 | 65.5 | |
| plan-and-executeModel=Qwen3-32B2026.01 | 73.6 | 78.1 | 53.8 | 63.3 | |
| AgriAgentModel=Qwen3-8B2026.01 | 70.3 | 76.8 | 67.2 | 58.9 | |
| AgriAgent (w/o feedback)Model=Qwen3-8B2026.01 | 69.3 | 77.1 | 69 | 60 | |
| reactModel=Qwen2.5-14B2026.01 | 69.3 | 74.6 | 37.8 | 51.3 | |
| plan-and-executeModel=Qwen2.5-7B2026.01 | 67 | 69.3 | 34.1 | 48.6 | |
| plan-and-executeModel=Qwen2.5-14B2026.01 | 64.9 | 76.9 | 37.5 | 53.4 | |
| AgriAgentModel=LLaMA3-8B2026.01 | 63.5 | 74.8 | 55.7 | 53.6 | |
| plan-and-executeModel=AgriGPT-8B2026.01 | 63 | 77.5 | 45.5 | 55.8 | |
| plan-and-executeModel=Qwen3-8B2026.01 | 61.2 | 76.1 | 45.1 | 54 | |
| reactModel=Qwen2.5-7B2026.01 | 58.9 | 66.7 | 29.7 | 43.2 | |
| reactModel=LLaMA3-8B2026.01 | 52.2 | 42.4 | 28.6 | 45.3 | |
| reactModel=AgriGPT-8B2026.01 | 50.5 | 71.7 | 32.6 | 42.3 | |
| AgriAgent (w/o feedback)Model=Qwen2.5-7B2026.01 | 50.4 | 70.5 | 47.8 | 41.8 | |
| reactModel=Qwen3-8B2026.01 | 49.6 | 73.8 | 31.5 | 44.8 | |
| reactModel=Qwen3-32B2026.01 | 47.5 | 70 | 30.1 | 44.6 | |
| AgriAgent (w/o feedback)Model=LLaMA3-8B2026.01 | 44.1 | 74.6 | 45.2 | 39.6 | |
| plan-and-executeModel=LLaMA3-8B2026.01 | 40 | 41.6 | 20 | 40 |