Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 812 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Question Answering | TriviaQA Gen (test) | 31 | Feb 26, 2026 | ||
| Math | OlympiadBench | 31 | Apr 20, 2026 | ||
| Reasoning | ZebraLogic | 31 | Apr 20, 2026 | ||
| Hyper-data Cleaning | MNIST (test) | 31 | Feb 26, 2026 | ||
| CUDA Kernel Generation | KernelBench Level 3 | 31 | Feb 26, 2026 | ||
| CUDA Kernel Generation | KernelBench Level 2 | 31 | Feb 26, 2026 | ||
| CUDA Kernel Generation | KernelBench Level 1 | 31 | Feb 26, 2026 | ||
| Next-item prediction | KuaiSAR | 31 | May 11, 2026 | ||
| Next-action prediction | OPeRA (test) | 31 | Apr 14, 2026 | ||
| Recommendation | Office | 31 | Mar 23, 2026 | ||
| Sequential Recommendation | MicroLens (test) | 31 | Mar 10, 2026 | ||
| Deepfake Detection | WildDeepfake (WDF) | 31 | May 11, 2026 | ||
| Novel-view Synthesis | WildRGB-D | 31 | Jul 1, 2026 | ||
| Video Temporal Grounding | ActivityNet TimeLens | 31 | May 22, 2026 | ||
| Video Temporal Grounding | Charades-TimeLens | 31 | May 22, 2026 | ||
| Multi-view Clustering | Reuters | 31 | Jun 4, 2026 | ||
| Multi-view Crowd Counting | CityStreet | 31 | Mar 26, 2026 | ||
| Fine-grained classification | Aircraft | 31 | Jun 8, 2026 | ||
| Fine-grained classification | Cars | 31 | Jun 8, 2026 | ||
| Fine-grained classification | Pets | 31 | Jun 8, 2026 | ||
| Fine-grained classification | Caltech101 | 31 | Jun 8, 2026 | ||
| Few-shot image classification | all-to-all setting | 31 | Feb 27, 2026 | ||
| Referring Segmentation | RISBench (test) | 31 | Mar 19, 2026 | ||
| Radar Nowcasting | SEVIR | 31 | Feb 26, 2026 | ||
| Deepfake Detection | Cross-Domain Evaluation (test) | 31 | Feb 26, 2026 |