Atari 100K atari-100k Leaderboard
Sample-efficient deep RL on the 26-game Atari 100K benchmark — agents train on only 100K environment steps (~2 hours of gameplay). Headline metric is the human-normalized median score across the 26 games (Human = 1.0). · Metric: Human-Normalized Median (higher is better)
| # | Model | Human-Normalized Median | Paper |
|---|---|---|---|
| 1 | EfficientZero V2 | 1.29 | — |
| 2 | EfficientZero | 1.09 | — |
| 3 | BBF (Bigger, Better, Faster) | 0.92 | — |
| 4 | DreamerV3 | 0.49 | — |
| 5 | SPR (Self-Predictive Representations) | 0.41 | — |
| 6 | DrQ | 0.27 | — |
| 7 | MuZero (Atari 100K) | 0.23 | — |
| 8 | CURL | 0.17 | — |
| 9 | SimPLe | 0.14 | — |