atari-2600
Unknown
OCR benchmark
9
Total Results
9
Models Tested
1
Metrics
2025-12-19
Last Updated
human-normalized-score
Higher is better
| Rank | Model | Score | Source |
|---|---|---|---|
| 1 | go-explore Exploration-focused agent. Score is Mean HNS (skewed by Montezuma's Revenge), not Median. | 40000 | nature-paper |
| 2 | agent57 Median HNS across 57 games. First to beat human baseline on ALL games. | 4731.3 | deepmind-research |
| 3 | bbos-1 Model-based optimization. | 1100 | research |
| 4 | gdi-h3 High sample efficiency. | 950 | research |
| 5 | dreamerv3 Mastered Atari with fixed hyperparameters using world models. | 840 | arxiv-paper |
| 6 | muzero Model-based agent planning with learned model. | 731 | nature-paper |
| 7 | rainbow-dqn Median HNS. Combines 7 improvements to DQN. | 231 | aaai-paper |
| 8 | human-gamer Professional human tester baseline. | 100 | baseline |
| 9 | dqn Historical baseline (2015). Median HNS. | 79 | nature-paper |