Deep Research on DeepResearch benchmark
53.4Average ScoreOur 5.1
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Our 5.1System Category=Deep Research Systems, Backbone Model=GPT-5.12026.04 | 53.4 | 54.85 | 55.63 | 50.82 | 49.84 | |
| Tavily ResearchSystem Category=Deep Research Systems2026.04 | 52.44 | 52.84 | 53.59 | 51.92 | 49.21 | |
| ThinkDepth.aiSystem Category=Deep Research Systems2026.04 | 52.43 | 52.02 | 53.88 | 52.04 | 50.12 | |
| GPT-5.1System Category=LLM with Public Search Tool, Backbone Model=GPT-5.12026.04 | 51.51 | 50.72 | 51.32 | 52.27 | 51.76 | |
| Salesforce AIRSystem Category=Deep Research Systems2026.04 | 50.65 | 50 | 51.09 | 50.77 | 50.32 | |
| Our 4.1System Category=Deep Research Systems, Backbone Model=GPT-4.12026.04 | 47.02 | 47.95 | 45.09 | 48.81 | 46.72 | |
| Our 4.1mSystem Category=Deep Research Systems, Backbone Model=GPT-4.1m2026.04 | 43.46 | 44.52 | 40.35 | 46.69 | 42.99 | |
| GPT-4.1System Category=LLM with Public Search Tool, Backbone Model=GPT-4.12026.04 | 29.31 | 25.59 | 18.42 | 40.63 | 36.49 |