Papers › Agent57: Outperforming the Atari Human Benchmark

Agent57: Outperforming the Atari Human Benchmark

30 Mar 2020ICML 2020 1arXiv:2003.13350archive 2025-07-28

Adrià Puigdomènech Badia, Bilal Piot, Steven Kapturowski, Pablo Sprechmann, Alex Vitvitskyi, Daniel Guo, Charles Blundell

Atari games have been a long-standing benchmark in the reinforcement learning (RL) community for the past decade. This benchmark was proposed to test general competency of RL algorithms. Previous work has achieved good average performance by doing outstandingly well on many games of the set, but very poorly in several of the most challenging games. We propose Agent57, the first deep RL agent that outperforms the standard human benchmark on all 57 Atari games. To achieve this result, we train a neural network which parameterizes a family of policies ranging from very exploratory to purely exploitative. We propose an adaptive mechanism to choose which policy to prioritize throughout the training process. Additionally, we utilize a novel parameterization of the architecture that allows for more consistent and stable learning.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 9 code samples harvested from 2 repositories linked to this paper: 8 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: community (archive-listed): 9 samples from 2 repositories, 8 ran (8 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 2 of its repositories that Syntology harvested for other papers:

Levels written on 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

Nkluge-correa/teeny-tiny_castle mentioned on GitHubtfApache-2.0Syntology: 3 samples, 3 ran (0 instrument-failed) · 0 unverified report
YHL04/agent57 mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Never Give Up: Learning Directed Exploration Strategies (arXiv:2002.06038; that paper's own run record: 7 ran (of which 3 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 2 where Syntology's instrument failed) · 7 unverified); nothing checks that they implement this paper's method. report
michaelnny/deep_rl_zoo mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Playing Atari with Deep Reinforcement Learning (arXiv:1312.5602; that paper's own run record: 64 ran (of which 24 constructed an object rather than computing a result; 46 with no instrument failure: 5 honoured, 0 violated, 41 with no contract checked; 18 where Syntology's instrument failed) · 53 unverified) and Never Give Up: Learning Directed Exploration Strategies (arXiv:2002.06038; that paper's own run record: 7 ran (of which 3 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 2 where Syntology's instrument failed) · 7 unverified); nothing checks that they implement this paper's method. report
yuta0821/agent57_pytorch mentioned on GitHubpytorchSyntology: not harvested report
pocokhc/agent57 tfMITSyntology: 6 samples, 5 ran (0 instrument-failed) · 1 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

9 samples harvested: 8 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

8ran
1unverified

Licence: 0 of the 9 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

Nkluge-correa/teeny-tiny_castle community (archive-listed) 3 samples, 3 ran (0 instrument-failed) · 0 unverified
create_dataframe in 7 papers ML-Intro-Course/20_xgboost_api.py ran Apache-2.0 (permissive) · 7e5fa12aac1ee278 · report fixture: signature list1:list/1/datetime, list2:list/1/float · returned DataFrame · level recorded 2026-09-24
create_sales_features ML-Intro-Course/20_xgboost_api.py ran Apache-2.0 (permissive) · 204108ac9fc9996f · report fixture: signature df:DataFrame/0/? · returned DataFrame · level recorded 2026-09-24
create_time_features ML-Intro-Course/20_xgboost_api.py ran Apache-2.0 (permissive) · 3aaed4f659384de1 · report fixture: signature df:DataFrame/0/? · returned DataFrame · level recorded 2026-09-24
pocokhc/agent57 community (archive-listed) 6 samples, 5 ran (0 instrument-failed) · 1 unverified
rank_sum agent/memory.py ran fingerprinted MIT (permissive) · 9b03860b997efb68 · report fixture: signature k:int/0/int, a:float/0/float · returned float · level recorded 2026-09-24
rank_sum_inverse agent/memory.py ran fingerprinted MIT (permissive) · caefea246b495774 · report fixture: signature k:int/0/int, a:float/0/float · returned float · level recorded 2026-09-24
rescaling agent/common.py ran fingerprinted MIT (permissive) · dad0ddc5efb4eece · report fixture: signature x:float/0/float, epsilon:float/0/float · returned float64 · level recorded 2026-09-24
rescaling_inverse agent/common.py ran fingerprinted MIT (permissive) · aa5ef26d6a8f198a · report fixture: signature x:float/0/float, epsilon:float/0/float · returned float64 · level recorded 2026-09-24
sigmoid in 216 papers
census: 36 implementations from 56 papers, 7 distinct outputs across 3 buckets, one shared input per bucket
agent/common.py ran fingerprinted MIT (permissive) · 65ee8b6d0e1288d5 · report fixture: signature x:ndarray/1/float, a:int/0/int · returned ndarray · level recorded 2026-09-24
clipped_error_loss agent/model.py unverified MIT (permissive) · 4a7f830ff4358d49 · report

Tasks

Atari GamesReinforcement LearningReinforcement Learning (RL)

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Atari Games Atari 2600 Alien Agent57 Score 297638.17 #2 of 49 Archive leaderboard report
Atari Games Atari 2600 Amidar Agent57 Score 29660.08 #1 of 48 Archive leaderboard report
Atari Games Atari 2600 Assault Agent57 Score 67212.67 #4 of 45 Archive leaderboard report
Atari Games Atari 2600 Asterix Agent57 Score 991384.42 #4 of 49 Archive leaderboard report
Atari Games Atari 2600 Asteroids Agent57 Score 150854.61 #8 of 43 Archive leaderboard report
Atari Games Atari 2600 Atlantis Agent57 Score 1528841.76 #7 of 42 Archive leaderboard report
Atari Games Atari 2600 Bank Heist Agent57 Score 23071.5 #3 of 45 Archive leaderboard report
Atari Games Atari 2600 Battle Zone Agent57 Score 934134.88 #1 of 47 Archive leaderboard report
Atari Games Atari 2600 Beam Rider Agent57 Score 300509.8 #4 of 49 Archive leaderboard report
Atari Games Atari 2600 Berzerk Agent57 Score 61507.83 #3 of 39 Archive leaderboard report
Atari Games Atari 2600 Bowling Agent57 Score 251.18 #3 of 44 Archive leaderboard report
Atari Games Atari 2600 Boxing Agent57 Score 100 #5 of 45 Archive leaderboard report
Atari Games Atari 2600 Breakout Agent57 Score 790.4 #9 of 58 Archive leaderboard report
Atari Games Atari 2600 Centipede Agent57 Score 412847.86 #6 of 45 Archive leaderboard report
Atari Games Atari 2600 Chopper Command Agent57 Score 999900 #4 of 45 Archive leaderboard report
Atari Games Atari 2600 Crazy Climber Agent57 Score 565909.85 #1 of 49 Archive leaderboard report
Atari Games Atari 2600 Defender Agent57 Score 677642.78 #6 of 21 Archive leaderboard report
Atari Games Atari 2600 Demon Attack Agent57 Score 143161.44 #7 of 46 Archive leaderboard report
Atari Games Atari 2600 Double Dunk Agent57 Score 23.93 #6 of 43 Archive leaderboard report
Atari Games Atari 2600 Enduro Agent57 Score 2367.71 #7 of 48 Archive leaderboard report
Atari Games Atari 2600 Fishing Derby Agent57 Score 86.97 #2 of 44 Archive leaderboard report
Atari Games Atari 2600 Freeway Agent57 Score 32.59 #24 of 59 Archive leaderboard report
Atari Games Atari 2600 Frostbite Agent57 Score 541280.88 #2 of 53 Archive leaderboard report
Atari Games Atari 2600 Gopher Agent57 Score 117777.08 #8 of 43 Archive leaderboard report
Atari Games Atari 2600 Gravitar Agent57 Score 19213.96 #1 of 53 Archive leaderboard report
Atari Games Atari 2600 HERO Agent57 Score 114736.26 #1 of 44 Archive leaderboard report
Atari Games Atari 2600 Ice Hockey Agent57 Score 63.64 #4 of 43 Archive leaderboard report
Atari Games Atari 2600 James Bond Agent57 Score 135784.96 #4 of 45 Archive leaderboard report
Atari Games Atari 2600 Kangaroo Agent57 Score 24034.16 #1 of 47 Archive leaderboard report
Atari Games Atari 2600 Krull Agent57 Score 251997.31 #3 of 45 Archive leaderboard report
Atari Games Atari 2600 Kung-Fu Master Agent57 Score 206845.82 #4 of 43 Archive leaderboard report
Atari Games Atari 2600 Montezuma's Revenge Agent57 Score 9352.01 #4 of 50 Archive leaderboard report
Atari Games Atari 2600 Ms. Pacman Agent57 Score 63994.44 #3 of 47 Archive leaderboard report
Atari Games Atari 2600 Name This Game Agent57 Score 54386.77 #4 of 43 Archive leaderboard report
Atari Games Atari 2600 Phoenix Agent57 Score 908264.15 #3 of 21 Archive leaderboard report
Atari Games Atari 2600 Pitfall! Agent57 Score 18756.01 #2 of 23 Archive leaderboard report
Atari Games Atari 2600 Pong Agent57 Score 20.67 #23 of 52 Archive leaderboard report
Atari Games Atari 2600 Private Eye Agent57 Score 79716.46 #2 of 52 Archive leaderboard report
Atari Games Atari 2600 Q*Bert Agent57 Score 580328.14 #1 of 57 Archive leaderboard report
Atari Games Atari 2600 River Raid Agent57 Score 63318.67 #4 of 41 Archive leaderboard report
Atari Games Atari 2600 Road Runner Agent57 Score 243025.8 #7 of 44 Archive leaderboard report
Atari Games Atari 2600 Robotank Agent57 Score 127.32 #2 of 42 Archive leaderboard report
Atari Games Atari 2600 Seaquest Agent57 Score 999997.63 #3 of 57 Archive leaderboard report
Atari Games Atari 2600 Skiing Agent57 Score -4202.6 #14 of 23 Archive leaderboard report
Atari Games Atari 2600 Solaris Agent57 Score 44199.93 #1 of 23 Archive leaderboard report
Atari Games Atari 2600 Space Invaders Agent57 Score 48680.86 #7 of 55 Archive leaderboard report
Atari Games Atari 2600 Star Gunner Agent57 Score 839573.53 #1 of 43 Archive leaderboard report
Atari Games Atari 2600 Surround Agent57 Score 9.5 #5 of 15 Archive leaderboard report
Atari Games Atari 2600 Tennis Agent57 Score 23.84 #5 of 43 Archive leaderboard report
Atari Games Atari 2600 Time Pilot Agent57 Score 405425.31 #5 of 44 Archive leaderboard report
Atari Games Atari 2600 Tutankham Agent57 Score 2354.91 #1 of 44 Archive leaderboard report
Atari Games Atari 2600 Up and Down Agent57 Score 623805.73 #7 of 44 Archive leaderboard report
Atari Games Atari 2600 Venture Agent57 Score 2623.71 #1 of 55 Archive leaderboard report
Atari Games Atari 2600 Video Pinball Agent57 Score 992340.74 #2 of 42 Archive leaderboard report
Atari Games Atari 2600 Wizard of Wor Agent57 Score 157306.41 #2 of 41 Archive leaderboard report
Atari Games Atari 2600 Yars Revenge Agent57 Score 998532.37 #1 of 17 Archive leaderboard report
Atari Games Atari 2600 Zaxxon Agent57 Score 249808.9 #2 of 41 Archive leaderboard report
Atari Games Atari games Agent57 Mean Human Normalized Score 4763.69% #5 of 12 Archive leaderboard report
Atari Games atari game Agent57 Human World Record Breakthrough 18 #3 of 9 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections