Question Answering on ZH-4O (test)
86.33QA AccuracyTELEMEM
Evaluation Results
| Method | Links | |
|---|---|---|
| TELEMEMBackbone LLM=Qwen3-8B (no-think mode), Embedding Model=Qwen3-8B-embedding2025.12 | 86.33 | |
| Long context LLMBackbone LLM=Qwen3-8B (no-think mode), Embedding Model=Qwen3-8B-embedding2025.12 | 84.92 | |
| MemobaseBackbone LLM=Qwen3-8B (no-think mode), Embedding Model=Qwen3-8B-embedding2025.12 | 76.78 | |
| A-MemBackbone LLM=Qwen3-8B (no-think mode), Embedding Model=Qwen3-8B-embedding2025.12 | 73.78 | |
| MOOMBackbone LLM=Qwen3-8B (no-think mode), Embedding Model=Qwen3-8B-embedding2025.12 | 72.6 | |
| Mem0Backbone LLM=Qwen3-8B (no-think mode), Embedding Model=Qwen3-8B-embedding2025.12 | 70.2 | |
| RAGBackbone LLM=Qwen3-8B (no-think mode), Embedding Model=Qwen3-8B-embedding2025.12 | 62.45 |