Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 2771 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Psychological defense mechanism classification | PSYDEFCONV (test) | 14 | Feb 26, 2026 | ||
| Verdict Prediction | AITA 1.0 (author split) | 14 | Feb 26, 2026 | ||
| Judgment Prediction | AITA verdict | 14 | Feb 26, 2026 | ||
| Task success | WebShop | 14 | Feb 27, 2026 | ||
| Task success | TextWorld | 14 | Feb 27, 2026 | ||
| Task success | SciWorld | 14 | Feb 27, 2026 | ||
| Task success | ALFWorld | 14 | Feb 27, 2026 | ||
| Faithfulness Hallucination Detection | LLM-AggreFact & HoVer Refined | 14 | Feb 26, 2026 | ||
| Multi-hop Faithfulness Hallucination Detection | HoVer Refined | 14 | Feb 26, 2026 | ||
| Faithfulness Hallucination Detection | LLM-AggreFact Refined | 14 | Feb 26, 2026 | ||
| Adversarial Attack Detection | Agentic AI workflow benchmark internal | 14 | Feb 26, 2026 | ||
| Safety Classification | HarmBench | 14 | Feb 26, 2026 | ||
| Safety Classification | AegisSafetyTest V2 | 14 | May 29, 2026 | ||
| Safety Classification | AegisSafety V1 (test) | 14 | Feb 26, 2026 | ||
| Safety Classification | XSTestResponse | 14 | Feb 26, 2026 | ||
| Safety Classification | Aya Redteaming | 14 | Feb 26, 2026 | ||
| Safety Classification | SimpleSafetyTests | 14 | Feb 26, 2026 | ||
| Machine Unlearning Evaluation | TOFU (Forget) | 14 | Feb 26, 2026 | ||
| Machine Unlearning Evaluation | TOFU Real World | 14 | Feb 26, 2026 | ||
| Machine Unlearning Evaluation | TOFU (Real Authors) | 14 | Feb 26, 2026 | ||
| Language Modeling | Pretraining Dataset | 14 | May 22, 2026 | ||
| Question Answering and Reasoning | Downstream Reasoning Suite (Arc-e, PIQA, Hellaswag, OpenBookQA, Winogrande, MMLU, BoolQ) | 14 | Feb 26, 2026 | ||
| General Reasoning | HellaSwag (test) | 14 | Mar 20, 2026 | ||
| Out-of-Domain (OOD) Bias Evaluation | Winobias | 14 | Feb 26, 2026 | ||
| Out-of-Domain (OOD) Bias Evaluation | Stereoset | 14 | Feb 26, 2026 |