Unverified50% confidenceReleaseExact time
Kijai updated the VAE model for MiniMax-H3 (MH3), adopting an int8 quantized version that significantly reduces VRAM usage with almost no loss in image quality.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/18/2026
First Seen
Valid until: 12/17/2026
Sources
Related Entities
Related Claims
UnverifiedThe updated VAE model filename is minimax_h3_video_vae_int8_convrot.safetensors.65% similarUnverifiedModel loading is divided into three categories: Clip model (select the int8 version), UNet large model (minimax_h3_int8 version), and VAE decoder.62% similarUnverifiedThe model trained a rank-256 LoRA adapter on top of the base model huihui-ai/Huihui-Qwen3.8-27B-abliterated.59% similarUnverifiedvLLM v0.27.0 adds support for the K-EXAONE-2.0-750B-A37B large-scale MoE model.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/931902API
curl https://kongchang.com/api/v1/knowledge/claims/931902MCP
get_claim(id=931902)