Retrieval-Augmented Question Answering on DeepSearch NQ
86SRSynapse
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SynapseTaskExecutor=Gemini-3.0-flash2026.04 | 86 | 53.6 | |
| SynapseTaskExecutor=gpt-4o-mini, Category=Baselines & Retrieval-Based Memory2026.04 | 52 | 38.6 | |
| SynapseTaskExecutor=gpt-4o2026.04 | 48 | 38.9 | |
| SynapseTaskExecutor=Gemini-2.5-flash2026.04 | 42 | 81.4 | |
| Gemini-3.0-flashTaskExecutor=gpt-4o-mini, Category=Zero-Shot Curators2026.04 | 34 | 8 | |
| ActiveContextTaskExecutor=Gemini-3.0-flash2026.04 | 32 | 6.8 | |
| Full ContextTaskExecutor=Gemini-3.0-flash2026.04 | 30 | 62.4 | |
| gpt-4oTaskExecutor=gpt-4o-mini, Category=Zero-Shot Curators2026.04 | 30 | 11.2 | |
| ActiveContextTaskExecutor=gpt-4o2026.04 | 28 | 7.2 | |
| Gemini-2.5-flashTaskExecutor=gpt-4o-mini, Category=Zero-Shot Curators2026.04 | 28 | 7.9 | |
| gpt-4o-miniTaskExecutor=gpt-4o-mini, Category=Zero-Shot Curators2026.04 | 28 | 11.2 | |
| ContextCuratorTaskExecutor=gpt-4o-mini, Category=RL-Trained Curator2026.04 | 28 | 7 | |
| ActiveContextTaskExecutor=Gemini-2.5-flash2026.04 | 27.9 | 8.1 | |
| Training-Free GRPOTaskExecutor=gpt-4o2026.04 | 26 | 35.3 | |
| AWMTaskExecutor=gpt-4o-mini, Category=Baselines & Retrieval-Based Memory2026.04 | 24 | 39.1 | |
| Mem0TaskExecutor=gpt-4o-mini, Category=Baselines & Retrieval-Based Memory2026.04 | 23 | 46.7 | |
| Training-Free GRPOTaskExecutor=Gemini-3.0-flash2026.04 | 22 | 78.2 | |
| Full ContextTaskExecutor=Gemini-2.5-flash2026.04 | 22 | 38.2 | |
| Full ContextTaskExecutor=gpt-4o2026.04 | 22 | 40.8 | |
| Qwen2.5-7B-InstructTaskExecutor=gpt-4o-mini, Category=Zero-Shot Curators2026.04 | 22 | 10.7 | |
| Full ContextTaskExecutor=gpt-4o-mini, Category=Baselines & Retrieval-Based Memory2026.04 | 20 | 35.5 | |
| Training-Free GRPOTaskExecutor=gpt-4o-mini, Category=Baselines & Retrieval-Based Memory2026.04 | 20 | 40.1 | |
| Training-Free GRPOTaskExecutor=Gemini-2.5-flash2026.04 | 10 | 81.2 |