Question Answering on NQ 10K context (Accuracy)
61.8AccuracyMemOCR
Evaluation Results
| Method | Links | |
|---|---|---|
| MemOCRMemory Budget (Tokens)=10242026.01 | 61.8 | |
| MemOCRMemory Budget (Tokens)=2562026.01 | 57.3 | |
| MemAgentMemory Budget (Tokens)=10242026.01 | 53.4 | |
| Qwen2.5-1MMemory Budget (Tokens)=100K2026.01 | 53.1 | |
| MemAgentMemory Budget (Tokens)=2562026.01 | 51.3 | |
| MemOCRMemory Budget (Tokens)=642026.01 | 51 | |
| Qwen2.5Memory Budget (Tokens)=100K2026.01 | 50.8 | |
| MemAgentMemory Budget (Tokens)=642026.01 | 48.7 | |
| Mem-αMemory Budget (Tokens)=2562026.01 | 47.9 | |
| Mem0Memory Budget (Tokens)=10242026.01 | 47.7 | |
| Mem0Memory Budget (Tokens)=2562026.01 | 46.9 | |
| Mem0Memory Budget (Tokens)=642026.01 | 46.9 | |
| Mem-αMemory Budget (Tokens)=10242026.01 | 45.6 | |
| MemOCRMemory Budget (Tokens)=162026.01 | 42.8 | |
| Mem0Memory Budget (Tokens)=162026.01 | 41.4 | |
| R1-Distill QwenMemory Budget (Tokens)=100K2026.01 | 35.2 | |
| Mem-αMemory Budget (Tokens)=642026.01 | 27.1 | |
| MemAgentMemory Budget (Tokens)=162026.01 | 24.5 | |
| Mem-αMemory Budget (Tokens)=162026.01 | 23.7 |