Agentic Writing on DeepResearchBench
49.6Pass@1Gemini 3-pro
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini 3-pro2026.03 | 49.6 | |
| GPT-5 High2026.03 | 48.7 | |
| Seed1.82026.03 | 43.5 | |
| Gemini 2.5-pro2026.03 | 40 | |
| Claude Sonnet-4.52026.03 | 39.9 |
| Method | Links | |
|---|---|---|
| Gemini 3-pro2026.03 | 49.6 | |
| GPT-5 High2026.03 | 48.7 | |
| Seed1.82026.03 | 43.5 | |
| Gemini 2.5-pro2026.03 | 40 | |
| Claude Sonnet-4.52026.03 | 39.9 |