Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1349 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| General Image Understanding | LLaVA-Bench In-the-Wild | 22 | Jun 2, 2026 | ||
| Agentic Task Completion | AppWorld (test-normal) | 22 | Jun 2, 2026 | ||
| Crystal Property Prediction | Materials Project (MP) | 22 | Jun 23, 2026 | ||
| AI-text detector attack effectiveness | RAID (evaluation) | 22 | Jun 2, 2026 | ||
| Hallucination Detection | WebQSP | 22 | Jun 2, 2026 | ||
| Question Answering | GridWorld (full) | 22 | Jun 2, 2026 | ||
| Interpretability and Faithfulness Evaluation | DINOv2 ViT-B/14 tokens | 22 | Jun 1, 2026 | ||
| Long-context language understanding | LongBench | 22 | Jun 1, 2026 | ||
| Image-Video Retrieval | ActivityNet | 22 | Jun 1, 2026 | ||
| Image Restoration | GGT-100K (test) | 22 | Jun 1, 2026 | ||
| Image Restoration | RealLQ social media, old photo, rain, haze, snow, and low light average (test) | 22 | Jun 1, 2026 | ||
| Text-to-Image Retrieval | ECCV Caption | 22 | Jun 1, 2026 | ||
| Visual-document retrieval | IRPAPERS | 22 | Jun 1, 2026 | ||
| Visual-document retrieval | VisDoc OOD | 22 | Jun 1, 2026 | ||
| Safety Evaluation | StrongReject | 22 | Jun 1, 2026 | ||
| Multi-Hop Search-augmented Question Answering | Bamboogle | 22 | Jun 11, 2026 | ||
| Multi-Hop Search-augmented Question Answering | MuSiQue | 22 | Jun 11, 2026 | ||
| Multi-Hop Search-augmented Question Answering | 2Wiki | 22 | Jun 11, 2026 | ||
| Multi-Hop Search-augmented Question Answering | HotpotQA | 22 | Jun 11, 2026 | ||
| Agentic Embodied Interactive Tasks | ALF Seen | 22 | Jun 1, 2026 | ||
| Memory Question Answering | LongMemEval | 22 | Jun 1, 2026 | ||
| Memory Question Answering | LoCoMo | 22 | Jun 1, 2026 | ||
| Automated Planning | IPC Planning Domains Various years | 22 | Jun 1, 2026 | ||
| Image Generation | CIFAR-10 50K 32x32 (test) | 22 | May 29, 2026 | ||
| Long-context reasoning | LongGenBench 8K | 22 | May 29, 2026 |