Scientific Discovery on ScienceAgentBench
0.225ScoreReAct
Evaluation Results
| Method | Links | |
|---|---|---|
| ReActBackbone=o4-mini2026.06 | 0.225 | |
| LATMBackbone=o4-mini2026.06 | 0.225 | |
| ReActBackbone=DeepSeek-V3.12026.06 | 0.206 | |
| LATMBackbone=DeepSeek-V3.12026.06 | 0.186 | |
| ML-Master2Backbone=o4-mini2026.06 | 0.176 | |
| ML-Master2Backbone=DeepSeek-V3.12026.06 | 0.157 | |
| ReActBackbone=GPT-4o2026.06 | 0.147 | |
| LAMBDABackbone=o4-mini2026.06 | 0.127 | |
| LAMBDABackbone=GPT-4o2026.06 | 0.118 | |
| EvoDS-evo-8BBackbone=Qwen3-8B2026.06 | 0.118 | |
| LAMBDABackbone=DeepSeek-V3.12026.06 | 0.108 | |
| LATMBackbone=GPT-4o2026.06 | 0.108 | |
| AutoGenBackbone=o4-mini2026.06 | 0.108 | |
| EvoDS-8BBackbone=Qwen3-8B2026.06 | 0.108 | |
| AutoGenBackbone=DeepSeek-V3.12026.06 | 0.098 | |
| Data InterpreterBackbone=o4-mini2026.06 | 0.098 | |
| Data InterpreterBackbone=DeepSeek-V3.12026.06 | 0.088 | |
| AutoGenBackbone=GPT-4o2026.06 | 0.088 | |
| Code InterpreterBackbone=GPT-4o2026.06 | 0.088 | |
| Data InterpreterBackbone=GPT-4o2026.06 | 0.088 | |
| ML-Master2Backbone=GPT-4o2026.06 | 0.088 | |
| DataMind-7BBackbone=Qwen2.5-Coder 7B2026.06 | 0.01 | |
| DataMind-14BBackbone=Qwen2.5-Coder 14B2026.06 | 0.01 | |
| DeepAnalyze-8BBackbone=DeepSeek-R1-0528-Qwen3-8B2026.06 | 0 |