Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 1637 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Multimodal Conversation | MMRole (ID) | 20 | Feb 26, 2026 | ||
| Analogy Solving | SAT Problems | 20 | Feb 27, 2026 | ||
| Analogy Solving | Green's Analogy Far-Analogy | 20 | Feb 27, 2026 | ||
| Analogy Solving | Green's Analogy (Near-Analogy) | 20 | Feb 27, 2026 | ||
| Step-level tool invocation safety detection | AgentDojo-Traj | 20 | Feb 26, 2026 | ||
| Step-level tool invocation safety detection | ASB-Traj | 20 | Feb 26, 2026 | ||
| Step-level tool invocation safety detection | AgentHarm Traj | 20 | Feb 26, 2026 | ||
| Story Completion | XStoryCloze | 20 | Feb 27, 2026 | ||
| Simultaneous Machine Translation | ACL60/60 En-De (eval) | 20 | Feb 26, 2026 | ||
| Machine Translation | Machine Translation Parallel Data (test) | 20 | Feb 26, 2026 | ||
| Knowledge Graph Question Answering | GrailQA (Overall) | 20 | Feb 26, 2026 | ||
| Watermark Detection | Ten-exam benchmark 1.0 (test) | 20 | Feb 26, 2026 | ||
| Watermarking Prevention | Ten-exam benchmark 1.0 (test) | 20 | Feb 26, 2026 | ||
| Cultural Pluralism Alignment (Cultural Values) | WVS (World Values Survey) | 20 | Feb 26, 2026 | ||
| Cultural Pluralism Alignment (Cultural Knowledge) | BLEND | 20 | Feb 26, 2026 | ||
| Visual Question Answering | MME | 20 | Jun 1, 2026 | ||
| Question Answering | TruthfulQA 1.0 (test) | 20 | Feb 26, 2026 | ||
| Mathematical Reasoning | AIME 25 | 20 | Mar 25, 2026 | ||
| Embodied Instruction Following | EB-ALFRED 1.0 (test) | 20 | Feb 26, 2026 | ||
| Causal Discovery | Alarm (d=37, |E|=46) medium-scale (test) | 20 | Feb 26, 2026 | ||
| Multimodal Future Prediction | FutureOmni 1.0 (Overall) | 20 | Feb 26, 2026 | ||
| Question Answering | BBQ (test) | 20 | Feb 26, 2026 | ||
| Multiple-choice reasoning | GPQA full dataset | 20 | Mar 25, 2026 | ||
| Long-horizon Question Answering | LoCoMo | 20 | Feb 26, 2026 | ||
| Long-horizon Question Answering | LoCoMo Overall All Categories 1.0 | 20 | Feb 26, 2026 |