HKU/ZJU/Fudan with Lightspeed launch HiSTEVE: hierarchical VLA agent playing Minecraft via native mouse and keyboard
tencent · hf · 2026-10-26
Researchers from HKU, Zhejiang University, Fudan and Lightspeed Studios released HiSTEVE 1.0 (arXiv, Oct 22, 2026), a hierarchical vision-language-action agent for open-world games.
- Decouples long-horizon planning from low-level execution in a layered architecture
- Acts through native keyboard-and-mouse control, mirroring the human player interface rather than abstract APIs
- Demonstrates complex open-ended task completion in Minecraft
The work shows VLA agents can operate real human-native interfaces in open-ended 3D environments.
More from Embodied
- CoRL paper: one-step inference matches multi-step for large VLAs, 1.6-4.8x faster action generation — kaixhin · 2026-10-26
- Reverse-engineered Ray-Ban Meta glasses show chess moves in AR at 3000+ Elo — Vjeux · 2026-10-26
- Hobbyist teaches StackChan robot to read scripts with Grok and Hermes-Agent — alexcovo_eth · 2026-10-26
- 10 wild humanoid robots: ex-Boston Dynamics droid butler, car-lifting bodyguard bot, ladder-climbing G1 — minchoi · 2026-10-26
- KAIST's humanoid v0.9 pushes 35 kg and lifts 20 kg, then moonwalks — minchoi · 2026-10-26
- Echo teaches VLA robots to detect task success and failure without frame-level labels — Yufeng Liu · 2026-10-26