Pith. sign in

REVIEW 5 cited by

Craftax: A Lightning-Fast Benchmark for Open-Ended Reinforcement Learning

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.16801 v2 pith:RUKTZSXN submitted 2024-02-26 cs.LG

classification cs.LG
keywords craftaxbenchmarkcrafterenvironmentlearningopen-endedbenchmarkschallenge
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Benchmarks play a crucial role in the development and analysis of reinforcement learning (RL) algorithms. We identify that existing benchmarks used for research into open-ended learning fall into one of two categories. Either they are too slow for meaningful research to be performed without enormous computational resources, like Crafter, NetHack and Minecraft, or they are not complex enough to pose a significant challenge, like Minigrid and Procgen. To remedy this, we first present Craftax-Classic: a ground-up rewrite of Crafter in JAX that runs up to 250x faster than the Python-native original. A run of PPO using 1 billion environment interactions finishes in under an hour using only a single GPU and averages 90% of the optimal reward. To provide a more compelling challenge we present the main Craftax benchmark, a significant extension of the Crafter mechanics with elements inspired from NetHack. Solving Craftax requires deep exploration, long term planning and memory, as well as continual adaptation to novel situations as more of the world is discovered. We show that existing methods including global and episodic exploration, as well as unsupervised environment design fail to make material progress on the benchmark. We believe that Craftax can for the first time allow researchers to experiment in a complex, open-ended environment with limited computational resources.

Discussion (0). Sign in to comment.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. DungeonBench: A Benchmark for Rules-Rich Tactical Reasoning in Dungeons & Dragons Combat

    cs.AI 2026-07 conditional novelty 7.0 of 10

    DungeonBench scores LLM tactical play on D&D combat, finding frontier policies clear ~80% of single encounters but only 40% of linked multi-encounter days.

  2. PACE: Parameter Change for Unsupervised Environment Design

    cs.LG 2026-05 unverdicted novelty 7.0 of 10

    PACE uses the squared L2 norm of policy parameter changes from a first-order approximation as an efficient proxy for environment value in UED, outperforming baselines with higher IQM and lower optimality gap on MiniGr...

  3. Automatic Generation of High-Performance RL Environments

    cs.LG 2026-03 conditional novelty 7.0 of 10

    Closed-loop prompt-based translation with hierarchical verification and iterative repair produces equivalent high-performance RL environments across five cases including new TCGJax.

  4. Diffusing Blame: Task-Dependent Credit Assignment in Biologically Plausible Dual-Stream Networks

    cs.LG 2026-06 unverdicted novelty 6.0 of 10

    Dual-stream excitatory/inhibitory networks trained with modulo error routing achieve 96.7% MNIST and 61.7% CIFAR-10 accuracy plus competitive RL performance, revealing task-dependent credit-assignment bottlenecks unde...

  5. stable-worldmodel: A Platform for Reproducible World Modeling Research and Evaluation

    cs.LG 2026-05 unverdicted novelty 5.0 of 10

    The paper presents stable-worldmodel (swm), a platform with high-performance data layer, modern world model baselines, planning solvers, and extended environments for reproducible research and generalization evaluation.

Pith tools