The authors built ASIWorld, a persistent survival environment where agent actions have delayed effects and earlier decisions shape later opportunities.
Survival sandbox benchmark evaluates autonomous agents on sustained goal execution
ASIWorld tests agents in a persistent environment with interdependent tasks and measures both outcomes and execution behaviors
Chinese Tech
Jianfeng Lin · Runze Wang · Haorong Hong · Yuxuan Xie · Zeno Zeng · Zhongyuan Liu
Lightspeed Studios, Tencent · The Hong Kong University of Science and Technology · University of Science and Technology of China · Hunyuan Team, Tencent
Research Digest··2 min read
Thread:Process Agent Benchmarks
The authors introduce ASIWorld, a survival sandbox with persistent world and character state, character-scoped observations, and a programmable asynchronous interface.
Why this paper
From Lightspeed Studios, Tencent and 3 others · Part of Process Agent Benchmarks, now 27 papers
In one line
ASIWorld is a survival sandbox benchmark that evaluates agents by task outcomes, execution behavior, and survival costs.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§