Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 2008 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Robotic Manipulation | PrimitiveSkill | 17 | May 6, 2026 | ||
| Embodied Navigation | Navigation | 17 | May 6, 2026 | ||
| Puzzle Reasoning | FrozenLake | 17 | May 6, 2026 | ||
| Mislabel Detection | apsfail | 17 | May 21, 2026 | ||
| Mislabel Detection | vehicle | 17 | May 21, 2026 | ||
| Mislabel Detection | pol | 17 | May 21, 2026 | ||
| Mislabel Detection | fraud | 17 | May 21, 2026 | ||
| Mislabel Detection | cpu | 17 | May 21, 2026 | ||
| Noisy Detection | wind | 17 | May 12, 2026 | ||
| Mislabel Detection | wind | 17 | May 21, 2026 | ||
| Mislabel Detection | 2dplanes | 17 | May 21, 2026 | ||
| Stealth Sycophancy Detection | Three-source Sycophancy Benchmark (test) | 17 | May 6, 2026 | ||
| Adversarial Generation | WOMD | 17 | May 20, 2026 | ||
| Jailbreak Attack | Llama 7B 2 | 17 | May 20, 2026 | ||
| Root Cause Analysis | AIOps Hard 2025 | 17 | May 6, 2026 | ||
| Root Cause Analysis | Fault injection Mid (unseen) | 17 | May 6, 2026 | ||
| Root Cause Analysis | SFT data Easy (seen) | 17 | May 6, 2026 | ||
| Question Answering | OpsEval Hard | 17 | May 6, 2026 | ||
| Question Answering | HITL-filtered Mid (unseen) | 17 | May 6, 2026 | ||
| Question Answering | SFT data Easy (seen) | 17 | May 6, 2026 | ||
| Classification | Orbit5k | 17 | May 6, 2026 | ||
| Perception | TreeBench | 17 | May 5, 2026 | ||
| Robotic Manipulation | SimplerEnv-WidowX | 17 | Jun 9, 2026 | ||
| Question Answering | PubMedQA long-context (PQA-L) | 17 | May 4, 2026 | ||
| Automated Student Response Grading | Beetle | 17 | May 4, 2026 |