Kimi K3 builds a VR companion in 30 minutes from one prompt and a 3D model
rohanpaul_ai · x · 2026-07-22
A Kimi staffer reportedly gave Kimi K3 a 3D model and a single prompt, and in about 30 minutes it built a VR companion that listens, replies, changes expressions, and moves across scenes.
- The model allegedly found and wired up local ASR and TTS, connected them to the LLM, and assembled the VR scenes on its own.
- The attached screenshots show the debugging/setup flow and a working VR scene, underscoring that this is a hands-on capability demo rather than a static benchmark claim.
Related event: Kimi K3 Creates VR Companion in 30 Minutes(3 posts)→
More from Embodied
- Snap Spectacles demo turns World Cup tracking data into editable photoreal replay — stspanho · 2026-07-22
- Embodied brain model trained on a 1,000-GPU cluster, says post — teortaxesTex · 2026-07-22
- Sam Altman and Jony Ive’s hardware project becomes part of OpenAI’s internal fight — 2C_ornot2C · 2026-07-22
- Robot manipulation needs mobile navigation and standardized benchmarks, researchers argue — YuXiang_IRVL · 2026-07-22
- Hikvision pitches physical AI with multimodal sensing and edge-cloud deployment — 智东西 · 2026-07-22
- Gaode upgrades ABot with five embodied AI models and claims SOTA on 17 benchmarks — 新智元 · 2026-07-22