DeepResearchBench — leaderboard
Evaluates deep research agents on multi-step information gathering and synthesis tasks. Tests ability to conduct thorough research across complex topics.
Metric: Average Score. Source: deepresearch-bench.github.io. Status: saturation imminent. 36 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 (High) | 55.31 |
| 2 | GPT-5 | 55.13 |
| 3 | Claude Sonnet 4.6 (High) | 54.87 |
| 4 | GPT-5.5 (High) | 54.01 |
| 5 | Claude Opus 4.6 (Medium) | 53.24 |
| 6 | GPT-5 (Low) | 51 |
| 7 | Claude Opus 4.8 (High) | 50.23 |
| 8 | Gemini 2.5 Pro | 49.71 |
| 9 | GPT-5 (Minimal) | 49.7 |
| 10 | Claude Opus 4.1 (20250805) | 49.7 |
| 11 | GPT-5.5 (Medium) | 49.55 |
| 12 | GPT-5 (Medium) | 49.5 |
| 13 | GPT-5.5 (Low) | 48.73 |
| 14 | GPT-5 (High) | 48.6 |
| 15 | Grok 4 | 47.9 |
Interactive version: theaggregate.ai/benchmark?slug=deepresearchbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.