General Purpose Tool Use on NESTFUL
96.3F1 FunctionalNo Memory
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| No MemoryBackbone=Claude 3.7 Sonnet2026.06 | 96.3 | 61.3 | 31.9 | 15.6 | 51.2 | |
| MemToolAgentBackbone=Claude 3.7 Sonnet2026.06 | 96 | 78 | 40.6 | 30.4 | 70.2 | |
| Extended ThinkingBackbone=Claude 3.7 Sonnet2026.06 | 95.2 | 62.7 | 33.4 | 16.9 | 66.1 | |
| ReAct AgentBackbone=Claude 3.7 Sonnet2026.06 | 93.2 | 56.4 | 29.5 | 10.8 | 58.2 |