Papers › Mastering Atari with Discrete World Models

Mastering Atari with Discrete World Models

5 Oct 2020ICLR 2021 1arXiv:2010.02193archive 2025-07-28

Danijar Hafner, Timothy Lillicrap, Mohammad Norouzi, Jimmy Ba

Intelligent agents need to generalize from past experience to achieve goals in complex environments. World models facilitate such generalization and allow learning behaviors from imagined outcomes to increase sample-efficiency. While learning world models from image inputs has recently become feasible for some tasks, modeling Atari games accurately enough to derive successful behaviors has remained an open challenge for many years. We introduce DreamerV2, a reinforcement learning agent that learns behaviors purely from predictions in the compact latent space of a powerful world model. The world model uses discrete representations and is trained separately from the policy. DreamerV2 constitutes the first agent that achieves human-level performance on the Atari benchmark of 55 tasks by learning behaviors inside a separately trained world model. With the same computational budget and wall-clock time, Dreamer V2 reaches 200M frames and surpasses the final performance of the top single-GPU agents IQN and Rainbow. DreamerV2 is also applicable to tasks with continuous actions, where it learns an accurate world model of a complex humanoid robot and solves stand-up and walking from only pixel inputs.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

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.

Syntology holds a reference implementation of DreamerV2, generated from the paper text (not the authors' code), at level 3 on Syntology's ladder: V2 “two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs” “neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)”; V3 “paper-derived property tests pass on both of those same two generations.” (quoted from Syntology's server card). For this implementation, level 3 is recorded, but its report lists no property test that passed on both. No rung compares this code against the authors' implementation. For agents: get_reference_implementation(name='DreamerV2') (how to connect); the free check have('2010.02193') needs no token.

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

Code

No sample from the archive's official repository (danijar/dreamerv2) is linked to this paper; the 15 samples below come from 3 community repositories. Syntology harvested code from danijar/dreamerv2 for another paper.

danijar/dreamerv2: Syntology's graph links these samples to Reward-Free Curricula for Training Robust World Models (arXiv:2306.09205; that paper's own run record: 5 ran (of which 1 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method.

Syntology 15 code samples harvested from 3 repositories linked to this paper: 11 ran (of which 1 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 1 violated, 9 with no contract checked; 1 where Syntology's instrument failed) · 4 unverified. Where Syntology's instrument failed: 1 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: community (archive-listed): 15 samples from 3 repositories, 11 ran (10 with no instrument failure; 1 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 4 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

danijar/dreamerv2 officialmentioned on GitHubtfMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Reward-Free Curricula for Training Robust World Models (arXiv:2306.09205; that paper's own run record: 5 ran (of which 1 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 1 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
CVC-Lab/SAC-for-H-Bond-Learning mentioned on GitHubpytorchSyntology: not harvested report
LukeBolly/dreamerv2 mentioned on GitHubtfMITSyntology: not harvested report
RajGhugare19/dreamerv2 mentioned on GitHubpytorchSyntology: 4 samples, 3 ran (1 instrument-failed) · 1 unverified report
adityabingi/Dreamer mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Dream to Control: Learning Behaviors by Latent Imagination (arXiv:1912.01603; that paper's own run record: 52 ran (of which 33 constructed an object rather than computing a result; 45 with no instrument failure: 1 honoured, 0 violated, 44 with no contract checked; 7 where Syntology's instrument failed) · 10 unverified); nothing checks that they implement this paper's method. report
andrejorsula/drl_grasping mentioned on GitHubpytorchBSD-3-ClauseSyntology: 3 samples, 2 ran (0 instrument-failed) · 1 unverified report
chandar-lab/LoCA2 mentioned on GitHubtfSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to The LoCA Regret: A Consistent Metric to Evaluate Model-Based Behavior in Reinforcement Learning (arXiv:2007.03158; that paper's own run record: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified; the one sample that ran constructed an object rather than computing a result); nothing checks that they implement this paper's method. report
iglu-contest/iglu-builder-baseline-dreamer mentioned on GitHubtfMITSyntology: 8 samples, 6 ran (0 instrument-failed) · 2 unverified report
opendilab/DI-engine pytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Hindsight Experience Replay (arXiv:1707.01495; that paper's own run record: 16 ran (of which 10 constructed an object rather than computing a result; 11 with no instrument failure: 1 honoured, 0 violated, 10 with no contract checked; 5 where Syntology's instrument failed) · 0 unverified), Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context (arXiv:1901.02860; that paper's own run record: 67 ran (of which 37 constructed an object rather than computing a result; 49 with no instrument failure: 4 honoured, 1 violated, 44 with no contract checked; 18 where Syntology's instrument failed) · 76 unverified), Decision Transformer: Reinforcement Learning via Sequence Modeling (arXiv:2106.01345; that paper's own run record: 17 ran (of which 10 constructed an object rather than computing a result; 12 with no instrument failure: 0 honoured, 0 violated, 12 with no contract checked; 5 where Syntology's instrument failed) · 9 unverified) and 1 more (every one in this page's JSON twin); nothing checks that they implement this paper's method. 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

15 samples harvested: 11 ran (of which 1 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 1 violated, 9 with no contract checked; 1 where Syntology's instrument failed) · 4 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.

1ran · violated contract
1ran · our draft was wrong
9ran
4unverified

Licence: 0 of the 15 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 3 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.

RajGhugare19/dreamerv2 community (archive-listed) 4 samples, 3 ran (1 instrument-failed) · 1 unverified
RSSMUtils dreamerv2/models/rssm.py ran MIT (permissive) · 945e8d12ec2f2ec5 · report fixture: returned RSSMUtils · level recorded 2026-09-22
batch_to_seq dreamerv2/models/rssm.py ran · our draft was wrong MIT (permissive) · 000455412d6f3ed9 · report fixture: signature batch_data:Tensor/3/float, batch_size:int/0/int, seq_len:int/0/int · returned Tensor · level recorded 2026-09-22
seq_to_batch dreamerv2/models/rssm.py ran · violated contract fingerprinted MIT (permissive) · 7dcf99ab88568073 · report fixture: signature sequence_data:Tensor/3/float, batch_size:int/0/int, seq_len:int/0/int · returned Tensor · level recorded 2026-09-22
RSSM in 7 papers dreamerv2/models/rssm.py unverified MIT (permissive) · 11fa7e68493b760d · report
andrejorsula/drl_grasping community (archive-listed) 3 samples, 2 ran (0 instrument-failed) · 1 unverified
flatten_dict_observations drl_grasping/utils/utils.py ran BSD-3-Clause (permissive) · 1f9296bbb07fb884 · report fixture: signature env:DictObsEnv/0/? · returned FlattenObservation · level recorded 2026-09-24
lowpass drl_grasping/utils/wrappers.py ran BSD-3-Clause (permissive) · aab451ef7c53d13e · report fixture: signature data:ndarray/1/float, freq:float/0/float, df:float/0/float, corners:int/0/int, zerophase:bool/0/bool · returned ndarray · level recorded 2026-09-24
get_wrapper_class drl_grasping/utils/utils.py unverified BSD-3-Clause (permissive) · 624c9428ed6a8179 · report
iglu-contest/iglu-builder-baseline-dreamer community (archive-listed) 8 samples, 6 ran (0 instrument-failed) · 2 unverified
flatten_conf common/other.py ran MIT (permissive) · db3fffc0e5f25954 · report fixture: signature config:dict/0/? · returned dict · level recorded 2026-09-24
load_baselines in 7 papers common/plot.py ran MIT (permissive) · b3ae56f77b905c0e · report fixture: returned list · level recorded 2026-09-24
load_episodes in 7 papers common/replay.py ran MIT (permissive) · 711ef26ab03f56dd · report fixture: signature directory:str/0/str, limit:NoneType/0/?, random:bool/0/bool · returned dict · level recorded 2026-09-24
save_episodes common/replay.py ran MIT (permissive) · 44b72fcd9872fef3 · report fixture: signature directory:str/0/str, episodes:list/1/dict · returned list · level recorded 2026-09-24
schedule in 32 papers common/other.py ran MIT (permissive) · c2659ae89b97624c · report fixture: signature string:str/0/str, step:EagerTensor/0/int · returned float · level recorded 2026-09-26
static_scan in 5 papers common/other.py ran MIT (permissive) · 77ff0f8ba58bad0e · report fixture: signature fn:callable/0/?, inputs:EagerTensor/2/float, start:EagerTensor/1/float, reverse:bool/0/bool · returned EagerTensor · level recorded 2026-09-26
load_run in 15 papers common/plot.py unverified MIT (permissive) · f90fbc5c43d007c0 · report
load_runs in 11 papers common/plot.py unverified MIT (permissive) · 4878c13e37288f19 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. For a generated assembly, an independent peer is a second implementation generated separately from the same paper text; neither implementation read the authors' code. At level 2 the two agreed numerically on the fixture's inputs; level 3 is recorded when property tests derived from the paper also passed on both, and each level 3 card prints how many its report lists. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it. A generated assembly's label was set when its code was generated for the method, and no model judged it; its card prints it as "by construction" or "not judged by a model". Where an assembly's phrase has no constructed figure, a brick that ran is not among the harvested samples this build reads, and whether a brick's run only constructed an object is read from that brick's harvested sample, so the figure was not counted; it does not say there were none.

2 assemblies on this paper: 1 harvested, 1 generated.

generated Assembly for DreamerV2 (graph method: DreamerV2), generated from the paper text, not the repository's code. 1 brick; individually on a fixture: 1 ran (1 with no instrument failure: 0 honoured, 1 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. Ran end-to-end against an independent peer; level 3 is recorded, but its report lists no property test that passed on both.

label matches by construction: the code was generated for this method (not judged by a model)

  1. straight_through_gradient entry · level 3 · ran · violated contract · 8a73fcddf4d99f2fb9993a7c0d02d87968a051488c673ab320fb4816093f79aa

harvested Assembly for DreamerV2 (graph method: DreamerV2). 4 bricks; individually on a fixture: 3 ran (of which 1 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 1 violated, 1 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified. The assembly has not been run end-to-end.

  1. RSSM entry · level 0 · unverified · 11fa7e68493b760d
  2. RSSMUtils helper · level 1 · ran · 945e8d12ec2f2ec5
  3. batch_to_seq helper · level 1 · ran · our draft was wrong · 000455412d6f3ed9
  4. seq_to_batch helper · level 1 · ran · violated contract · 7dcf99ab88568073

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2010.02193") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Atari Games

1 archive task tag without a task page not shown.

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Atari Games Atari 2600 Alien DreamerV2 Score 3967 #22 of 49 Archive leaderboard report
Atari Games Atari 2600 Amidar DreamerV2 Score 2577 #8 of 48 Archive leaderboard report
Atari Games Atari 2600 Assault DreamerV2 Score 23625 #9 of 45 Archive leaderboard report
Atari Games Atari 2600 Asterix DreamerV2 Score 72311 #20 of 49 Archive leaderboard report
Atari Games Atari 2600 Asteroids DreamerV2 Score 41526 #11 of 43 Archive leaderboard report
Atari Games Atari 2600 Atlantis DreamerV2 Score 978778 #12 of 42 Archive leaderboard report
Atari Games Atari 2600 Bank Heist DreamerV2 Score 1126 #21 of 45 Archive leaderboard report
Atari Games Atari 2600 Battle Zone DreamerV2 Score 40325 #14 of 47 Archive leaderboard report
Atari Games Atari 2600 Beam Rider DreamerV2 Score 18646 #22 of 49 Archive leaderboard report
Atari Games Atari 2600 Berzerk DreamerV2 Score 810 #33 of 39 Archive leaderboard report
Atari Games Atari 2600 Bowling DreamerV2 Score 49 #32 of 44 Archive leaderboard report
Atari Games Atari 2600 Boxing DreamerV2 Score 92 #28 of 45 Archive leaderboard report
Atari Games Atari 2600 Breakout DreamerV2 Score 312 #43 of 58 Archive leaderboard report
Atari Games Atari 2600 Centipede DreamerV2 Score 11883 #16 of 45 Archive leaderboard report
Atari Games Atari 2600 Chopper Command DreamerV2 Score 2861 #41 of 45 Archive leaderboard report
Atari Games Atari 2600 Crazy Climber DreamerV2 Score 161839 #15 of 49 Archive leaderboard report
Atari Games Atari 2600 Demon Attack DreamerV2 Score 82263 #21 of 46 Archive leaderboard report
Atari Games Atari 2600 Double Dunk DreamerV2 Score 17 #14 of 43 Archive leaderboard report
Atari Games Atari 2600 Enduro DreamerV2 Score 1656 #23 of 48 Archive leaderboard report
Atari Games Atari 2600 Fishing Derby DreamerV2 Score 65 #5 of 44 Archive leaderboard report
Atari Games Atari 2600 Freeway DreamerV2 Score 33 #21 of 59 Archive leaderboard report
Atari Games Atari 2600 Frostbite DreamerV2 Score 11384 #6 of 53 Archive leaderboard report
Atari Games Atari 2600 Gopher DreamerV2 Score 92282 #13 of 43 Archive leaderboard report
Atari Games Atari 2600 Gravitar DreamerV2 Score 3789 #12 of 53 Archive leaderboard report
Atari Games Atari 2600 HERO DreamerV2 Score 21868 #23 of 44 Archive leaderboard report
Atari Games Atari 2600 Ice Hockey DreamerV2 Score 26 #10 of 43 Archive leaderboard report
Atari Games Atari 2600 James Bond DreamerV2 Score 40445 #7 of 45 Archive leaderboard report
Atari Games Atari 2600 Kangaroo DreamerV2 Score 14064 #14 of 47 Archive leaderboard report
Atari Games Atari 2600 Krull DreamerV2 Score 50061 #8 of 45 Archive leaderboard report
Atari Games Atari 2600 Kung-Fu Master DreamerV2 Score 62741 #14 of 43 Archive leaderboard report
Atari Games Atari 2600 Montezuma's Revenge DreamerV2 Score 81 #28 of 50 Archive leaderboard report
Atari Games Atari 2600 Ms. Pacman DreamerV2 Score 5652 #18 of 47 Archive leaderboard report
Atari Games Atari 2600 Name This Game DreamerV2 Score 14649 #18 of 43 Archive leaderboard report
Atari Games Atari 2600 Phoenix DreamerV2 Score 49375 #15 of 21 Archive leaderboard report
Atari Games Atari 2600 Pitfall! DreamerV2 Score 0 #12 of 23 Archive leaderboard report
Atari Games Atari 2600 Pong DreamerV2 Score 20 #30 of 52 Archive leaderboard report
Atari Games Atari 2600 Private Eye DreamerV2 Score 2198 #18 of 52 Archive leaderboard report
Atari Games Atari 2600 Q*Bert DreamerV2 Score 94688 #8 of 57 Archive leaderboard report
Atari Games Atari 2600 River Raid DreamerV2 Score 16351 #19 of 41 Archive leaderboard report
Atari Games Atari 2600 Road Runner DreamerV2 Score 203576 #10 of 44 Archive leaderboard report
Atari Games Atari 2600 Robotank DreamerV2 Score 78 #8 of 42 Archive leaderboard report
Atari Games Atari 2600 Seaquest DreamerV2 Score 7480 #26 of 57 Archive leaderboard report
Atari Games Atari 2600 Skiing DreamerV2 Score -9299 #15 of 23 Archive leaderboard report
Atari Games Atari 2600 Solaris DreamerV2 Score 922 #22 of 23 Archive leaderboard report
Atari Games Atari 2600 Space Invaders DreamerV2 Score 2474 #35 of 55 Archive leaderboard report
Atari Games Atari 2600 Star Gunner DreamerV2 Score 7800 #37 of 43 Archive leaderboard report
Atari Games Atari 2600 Tennis DreamerV2 Score 14 #12 of 43 Archive leaderboard report
Atari Games Atari 2600 Time Pilot DreamerV2 Score 37945 #11 of 44 Archive leaderboard report
Atari Games Atari 2600 Tutankham DreamerV2 Score 264 #15 of 44 Archive leaderboard report
Atari Games Atari 2600 Up and Down DreamerV2 Score 653662 #5 of 44 Archive leaderboard report
Atari Games Atari 2600 Venture DreamerV2 Score 2 #46 of 55 Archive leaderboard report
Atari Games Atari 2600 Video Pinball DreamerV2 Score 41860 #37 of 42 Archive leaderboard report
Atari Games Atari 2600 Wizard of Wor DreamerV2 Score 12851 #16 of 41 Archive leaderboard report
Atari Games Atari 2600 Yars Revenge DreamerV2 Score 156748 #8 of 17 Archive leaderboard report
Atari Games Atari 2600 Zaxxon DreamerV2 Score 50699 #7 of 41 Archive leaderboard report
Atari Games Atari games DreamerV2 Mean Human Normalized Score 631.17% #11 of 12 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