Pending benchmarksTracking since Oct 19, 2026
Is XGEN-JING state of the art?Pending benchmarks.Announced Oct 14, 2026; no independent results yet.
XGEN Labs' egocentric interactive experience model on MiniMax-H3: first-person video and audio from actions, reference images and observation history.
Announced facts
The details so far
- Lab
- XGEN Labs
- Announced
- Oct 14, 2026
- API
- Not yet
- Open weights
- Published
- Type
- Egocentric interactive experience / world model
- Base
- MiniMaxAI/MiniMax-H3
- Release
- JING-Flash-v1 four-step bidirectional inference, code, examples and prompt skills; causal model listed as coming soon
- Inputs
- Actions, reference images and observation history
- Outputs
- First-person video and audio (navigation, object interaction, conversation)
- License
- MiniMax-H3 Community License Agreement (card license: other)
- Paper
- arXiv 2610.19726 (tech report)
Tags marked unsure are not stated on the lab's own pages.
Next steps
What would let us call it
Verdicts use only independent boards we already track. Two consecutive snapshots are required before any crown.
- Publicly announced with a source
- Usable by the public (API or weights)
- Listed on an independent board we track
- Enough results for a rank in a category
- Holds across two snapshots (crowns only)
Receipts
Sources
LABXGEN Labs announcement on Xx.com/XGEN_labs/status/2112188039644758509 · Oct 19, 2026PAPERXGEN-JING tech report (arXiv 2610.19726)arxiv.org/abs/2610.19726 · Oct 14, 2026REPOXGEN-JING on Hugging Facehuggingface.co/XGENlabs/XGEN-JING · Oct 19, 2026REPOXGEN-Labs/XGEN-JING on GitHubgithub.com/XGEN-Labs/XGEN-JING · Oct 19, 2026Tracking continues until an independent board lists it.