Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1994 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Entity-Swap Question Answering | NQ-Swap | 17 | Jun 4, 2026 | ||
| Open-Book Question Answering | NQ-Open Open-Book | 17 | Jun 4, 2026 | ||
| Abstractive Summarization Faithfulness | XSum | 17 | Jun 4, 2026 | ||
| Live Streaming Risk Assessment | June OOD (10/16–10/17) (test) | 17 | Jun 4, 2026 | ||
| Live Streaming Risk Assessment | May OOD 09/23–09/24 (test) | 17 | Jun 4, 2026 | ||
| Inference Cost Attack | HotpotQA | 17 | Jun 4, 2026 | ||
| Video Reasoning | VBVR-Bench Overall | 17 | Jun 2, 2026 | ||
| Complex Reasoning on Structured Data | WikiSQL-E | 17 | Jun 2, 2026 | ||
| Protein-protein interaction site prediction | 60 (test) | 17 | Jun 2, 2026 | ||
| Knowledge & Instruction Following | IFEval | 17 | Jun 2, 2026 | ||
| Knowledge & Instruction Following | MMLU-Pro | 17 | Jun 2, 2026 | ||
| Speech Enhancement | Oldenburg (N+V) | 17 | Jun 2, 2026 | ||
| Speech Enhancement | Oldenburg (V) | 17 | Jun 2, 2026 | ||
| Speech Enhancement | Oldenburg | 17 | Jun 2, 2026 | ||
| Harmful prompt detection | Combined Average | 17 | Jun 2, 2026 | ||
| Harmful prompt detection | WJB | 17 | Jun 2, 2026 | ||
| Harmful prompt detection | WGMix | 17 | Jun 2, 2026 | ||
| Harmful prompt detection | TChat | 17 | Jun 2, 2026 | ||
| Survival Analysis | TCGA-LUAD | 17 | Jun 4, 2026 | ||
| Visual Question Answering | GQA v1 (test) | 17 | Jun 2, 2026 | ||
| Comprehensive LLM Evaluation | PostTrainBench (test) | 17 | Jun 2, 2026 | ||
| General Reasoning Average | PostTrainBench | 17 | Jun 2, 2026 | ||
| Medical and Health Knowledge | HealthBench | 17 | Jun 2, 2026 | ||
| Function Calling | BFCL | 17 | Jun 2, 2026 | ||
| LLM Chat Evaluation | ArenaHard | 17 | Jun 2, 2026 |