Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 2598 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Medical Visual Question Answering | Slake 12 | 14 | Jun 19, 2026 | ||
| Robotic Manipulation | RoboTwin-MeM 1.0 (test) | 14 | Jun 19, 2026 | ||
| Light Field Occlusion Removal | 4-Syn Average | 14 | Jun 19, 2026 | ||
| Reward Modeling | PKU-SafeRLHF | 14 | Jun 19, 2026 | ||
| Reward Modeling | UltraFeedback | 14 | Jun 19, 2026 | ||
| Reward Modeling | HelpSteer | 14 | Jun 19, 2026 | ||
| Protocol performance evaluation | G-Lox Microbenchmark | 14 | Jun 19, 2026 | ||
| Language-conditioned robotic manipulation | LIBERO | 14 | Jun 19, 2026 | ||
| Commonsense Reasoning | StrategyQA | 14 | Jun 19, 2026 | ||
| Clinical Image Classification | 7 clinical datasets aggregate (CheXpert, COVID-BLUES, VinDr-Mammo, ISIC-2020, HAM10000, PAD-UFES-20, Hemorrhage) (test) | 14 | Jun 19, 2026 | ||
| Time Series Forecasting | Fund2 horizon 10 | 14 | Jun 19, 2026 | ||
| Time Series Forecasting | Fund1 horizon 10 | 14 | Jun 19, 2026 | ||
| Consistent Reasoning | GSM8K Verification | 14 | Jun 18, 2026 | ||
| Consistent Reasoning | TruthfulQA | 14 | Jun 18, 2026 | ||
| Natural scene image inpainting | Places2 (Small) | 14 | Jun 18, 2026 | ||
| Natural scene image inpainting | Places2 Large | 14 | Jun 18, 2026 | ||
| Drug-resistance variant scoring | Kim EGFR | 14 | Jun 18, 2026 | ||
| Drug-resistance variant scoring | Coelho 10g | 14 | Jun 18, 2026 | ||
| Multi-skill Compositional Reasoning | MM-Vet | 14 | Jun 18, 2026 | ||
| Multimodal Evaluation | MME | 14 | Jun 18, 2026 | ||
| OCR-based Visual Question Answering | TextVQA | 14 | Jun 18, 2026 | ||
| Visual Question Answering | GQA | 14 | Jun 18, 2026 | ||
| Neural Architecture Search | NAS-Bench-201 ImageNet16-120 (test) | 14 | Jun 18, 2026 | ||
| Pragmatic Reasoning | ALTPRAG | 14 | Jun 18, 2026 | ||
| A4 divergence attack defense | Real-LLM 8 targets × 3 phrasings full stack late 2025/2026 (val) | 14 | Jun 18, 2026 |