Voyager Minecraft

Voyager Minecraft: Measures embodied-agent, navigation, manipulation, or simulated robotics task success.

6rows
downstream_success_rateprimary metric
2026-05-27sampled

Metadata

Metrics

Downstream Success Rate, Successful Trials, Average Iterations on Solved Tasks (lower is better)

Latest Results

Rows are imported from the public Voyager arXiv source downstream generalization table.

Rank Subject Downstream Success Rate Model Match Provenance Sampled
1 Voyager (Ours) 100% Imported 2026-05-27
2 Voyager w/o Skill Library 91.6667% Imported 2026-05-27
3 AutoGPT w/ Voyager Skill Library 33.3333% Imported 2026-05-27
4 AutoGPT 0% Imported 2026-05-27
5 ReAct 0% Imported 2026-05-27
6 Reflexion 0% Imported 2026-05-27