Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1251 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| General Problem Solving | Mixed (AIME, GPQA, HLE, HotpotQA, ALFWorld) | 24 | Feb 26, 2026 | ||
| Humanities Question Answering | HLE | 24 | Feb 26, 2026 | ||
| Multimodal Fact-Level Attribution | Video-MMMU 1.0 (sampled examples) | 24 | Feb 26, 2026 | ||
| Multimodal Fact-Level Attribution | WorldSense 1.0 (sampled examples) | 24 | Feb 26, 2026 | ||
| Multi-Turn Function Calling | BFCL Multi-Turn Base v3 | 24 | Apr 20, 2026 | ||
| Selective Prediction | TriviaQA (test) | 24 | Feb 26, 2026 | ||
| Text-based safety moderation | Toxic Chat | 24 | Feb 26, 2026 | ||
| Orienteering Problem | OP | 24 | Jun 4, 2026 | ||
| Mathematical Reasoning | Minerva | 24 | Mar 25, 2026 | ||
| Mathematical Reasoning | AIME 24 | 24 | Mar 4, 2026 | ||
| Knowledge Tracing | DBE-KT22 | 24 | Apr 14, 2026 | ||
| Visual Reasoning | HalluBench | 24 | Feb 26, 2026 | ||
| Multi-hop Question Answering | 2WikiMultiHopQA N=200 | 24 | Feb 26, 2026 | ||
| Tool Use | Synthetic Data (test) | 24 | Feb 26, 2026 | ||
| Mathematical Reasoning | AIME 24 | 24 | Feb 26, 2026 | ||
| Deep Search | Browsecomp | 24 | May 29, 2026 | ||
| Hallucination Detection | Mixture-domain TriviaQA, CSQA, Belebele, CoQA, Math, SVAMP | 24 | Feb 26, 2026 | ||
| Code Localization | SWE-bench Verified (test) | 24 | Feb 26, 2026 | ||
| Question Answering | STRATEGYQA | 24 | Feb 26, 2026 | ||
| Adversarial Risk Estimation | HarmBench (test) | 24 | Feb 26, 2026 | ||
| General Reasoning | MMLU-P | 24 | Jun 1, 2026 | ||
| Over-refusal Assessment | XS (test) | 24 | Feb 26, 2026 | ||
| Jailbreak Attack Defense | FORTRESS | 24 | Feb 26, 2026 | ||
| Jailbreak Attack Defense | PAIR | 24 | May 7, 2026 | ||
| Jailbreak Attack Defense | GCG | 24 | May 7, 2026 |