Agent Memory on LoCoMo
1,365.6Token CountSummary
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Summarycompression_model=Gemini 3.1 Pro, answering_model=Gemini 3.1 Pro, evaluator_model=GPT-4o-mini2026.06 | 1,365.6 | 61.05 | 94.2 | |
| BabelTelecompression_model=Gemini 3.1 Pro, answering_model=Gemini 3.1 Pro, evaluator_model=GPT-4o-mini2026.06 | 1,382.2 | 62.53 | 96.48 | |
| Originalcompression_model=Gemini 3.1 Pro, answering_model=Gemini 3.1 Pro, evaluator_model=GPT-4o-mini2026.06 | 2,819.5 | 64.81 | 100 |