Benchmarks
Deepresearchbench
Models scored
36
Best score
0.55
Metric
Average score
Deepresearchbench is an AI evaluation tracked by Epoch AI, with 36 scored model versions on record. The scores here are reported by third parties rather than produced by Epoch AI running the evaluation itself.
Scored models span January 2025 to May 2026. The best score recorded is 0.55, measured as Average score. Models come from Anthropic, OpenAI, Google DeepMind, xAI, Perplexity, Google, DeepSeek.
Full record
- Score metric
- Average score
- Models scored
- 36
- Best score recorded
- 0.55
- Earliest model scored
- Jan 21, 2025
- Most recent model scored
- May 28, 2026
- Third-party reported
- Yes
Leaderboard
| 01 | claude-opus-4-6_high | Anthropic | 0.55 |
| 02 | gpt-5-2025-08-07_unknown | OpenAI | 0.55 |
| 03 | xiaoyi-deepresearch | 0.55 | |
| 04 | claude-sonnet-4-6_high | Anthropic | 0.55 |
| 05 | gpt-5.5_high | OpenAI | 0.54 |
| 06 | claude-opus-4-6_medium | Anthropic | 0.53 |
| 07 | claude-sonnet-4-5-20250929_2K | Anthropic | 0.53 |
| 08 | gpt-5-2025-08-07_low | OpenAI | 0.51 |
| 09 | claude-opus-4-8_high | Anthropic | 0.5 |
| 10 | gemini-2.5-pro | Google DeepMind | 0.5 |
| 11 | claude-opus-4-1-20250805 | Anthropic | 0.5 |
| 12 | gpt-5-2025-08-07_minimal | OpenAI | 0.5 |
| 13 | gpt-5.5_medium | OpenAI | 0.5 |
| 14 | gpt-5-2025-08-07_medium | OpenAI | 0.49 |
| 15 | claude-opus-4-8_low | Anthropic | 0.49 |
| 16 | claude-opus-4-20250514_2K | Anthropic | 0.49 |
| 17 | gpt-5.5_low | OpenAI | 0.49 |
| 18 | gpt-5-2025-08-07_high | OpenAI | 0.49 |
| 19 | grok-4-0709 | xAI | 0.48 |
| 20 | claude-sonnet-4-20250514_2K | Anthropic | 0.48 |
| 21 | claude-opus-4-8_medium | Anthropic | 0.47 |
| 22 | o3-2025-04-16_medium | OpenAI | 0.47 |
| 23 | claude-3-7-sonnet-20250219_2K | Anthropic | 0.44 |
| 24 | gemini-2.5-pro-preview-06-05 | Google DeepMind | 0.43 |
| 25 | gpt-5.1-2025-11-13_low | OpenAI | 0.43 |
| 26 | gpt-5.2-2025-12-11_low | OpenAI | 0.41 |
| 27 | sonar-pro | Perplexity | 0.38 |
| 28 | sonar | 0.38 | |
| 29 | claude-3-7-sonnet-20250219 | Anthropic | 0.37 |
| 30 | gemini-3.1-flash-lite | 0.36 | |
| 31 | gpt-5.4-mini-2026-03-17_low | OpenAI | 0.36 |
| 32 | DeepSeek-R1-0528 | DeepSeek | 0.35 |
| 33 | gpt-5.4-2026-03-05_low | OpenAI | 0.35 |
| 34 | gemini-2.5-pro-preview-05-06 | Google DeepMind | 0.32 |
| 35 | gpt-4.1-2025-04-14 | OpenAI | 0.29 |
| 36 | gemini-2.5-flash-preview-04-17 | Google DeepMind | 0.29 |
More benchmarks
SourceEpoch AI, 'AI Benchmarking Hub'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/benchmarks. Licensed under CC BY 4.0.
← All benchmarks