Research
Benchmarks
Compare state-of-the-art methods across the tasks and datasets used to measure AI progress.
309,131 benchmarksPage 930 of 12,366
Benchmarks
| Task | Dataset | Domain | Trend | Results | Last update |
|---|---|---|---|---|---|
| Scientific Reasoning | GPQA main | 28 | Jun 5, 2026 | ||
| Complex Reasoning | BBH | 28 | Jun 5, 2026 | ||
| Image Restoration | Cameraman gray scale image | 28 | Jun 5, 2026 | ||
| Active Object Detection | Handlebottle Simulation | 28 | Jun 5, 2026 | ||
| Image Classification | SynDigits-LT (gamma=200) | 28 | Jun 4, 2026 | ||
| Chess Outcome Prediction | Lichess sample April 2026 (month-wide slice) | 28 | Jun 4, 2026 | ||
| Retrieval-Augmented Generation | 2WikiMultiHopQA | 28 | Jun 4, 2026 | ||
| Uncertainty Calibration | TQA | 28 | Jun 4, 2026 | ||
| Image Reconstruction | FFHQ | 28 | Jun 4, 2026 | ||
| Mathematical Reasoning | U-Shaped Stream | 28 | Jun 4, 2026 | ||
| Mathematical Reasoning | Stream Mostly-Hard | 28 | Jun 4, 2026 | ||
| Mathematical Reasoning | Mostly-Easy Stream | 28 | Jun 4, 2026 | ||
| Mathematical Reasoning | Balanced Stream | 28 | Jun 4, 2026 | ||
| Science Simulation Task Completion | ScienceWorld Unseen | 28 | Jun 2, 2026 | ||
| Science Simulation Task Completion | ScienceWorld Seen | 28 | Jun 2, 2026 | ||
| Mathematical reasoning | MATH500 | 28 | Jun 30, 2026 | ||
| LLM Moderation | WildGuardMix (test) | 28 | Jun 2, 2026 | ||
| Robot Motion Planning | Semi-circular tunnel experiment | 28 | Jun 2, 2026 | ||
| Quantitative Reasoning | AIME | 28 | Jun 2, 2026 | ||
| Abstention Metacognition | AbstentionBench Normal Prompt | 28 | Jun 2, 2026 | ||
| Abstention Metacognition | AbstentionBench Abstention Prompt | 28 | Jun 2, 2026 | ||
| Mathematical Reasoning | Math tasks | 28 | Jun 2, 2026 | ||
| Image Classification | CelebA | 28 | Jul 1, 2026 | ||
| LLM Inference | Qwen3-8B 2k prompts Balanced workload | 28 | Jun 2, 2026 | ||
| Regression | Crime (USA) | 28 | Jun 2, 2026 |