Deep research on GALA
51.9Overall ScoreAgentDisCo w/ Rednote & Harness
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| AgentDisCo w/ Rednote & HarnessBackbone=Gemini-2.5-Pro, Retrieval source=Rednote, Harness optimization=true, Evaluation protocol=RACE2026.05 | 51.9 | 51.61 | 53.44 | 51.78 | 50.67 | |
| AgentDisCo w/ RednoteBackbone=Gemini-2.5-Pro, Retrieval source=Rednote, Evaluation protocol=RACE2026.05 | 51.02 | 50.88 | 51.11 | 51.25 | 50.95 | |
| AgentDisCo w/ Rednote & GoogleBackbone=Gemini-2.5-Pro, Retrieval source=Rednote & Google, Evaluation protocol=RACE2026.05 | 50.95 | 51.21 | 50.44 | 50.78 | 49.79 | |
| AgentDisCo w/ HarnessBackbone=Gemini-2.5-Pro, Harness optimization=true, Evaluation protocol=RACE2026.05 | 50.58 | 50.41 | 51.24 | 50.16 | 49.85 | |
| AgentDisCoBackbone=Gemini-2.5-Pro, Evaluation protocol=RACE2026.05 | 50 | 50 | 50 | 50 | 50 | |
| Doubao-ResearchVersion/Date=2026-04, Evaluation protocol=RACE2026.05 | 49.82 | 50.87 | 47.42 | 50.65 | 50.86 | |
| Qwen-ResearchVersion/Date=2026-04, Evaluation protocol=RACE2026.05 | 46.69 | 45.38 | 45.36 | 47.23 | 49.56 | |
| OpenAI o3-DeepResearchVersion/Date=2026-04, Evaluation protocol=RACE2026.05 | 45.88 | 45.37 | 42.88 | 48.04 | 47.72 |