Expired50% confidenceReleaseExact time
Minimax H3 now supports int8 quantized model variants.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
9/27/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/11/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedThe MiniMax H3 model uses a Diffusion Transformer (DiT) architecture and can run a complete inference pipeline on a local GPU.76% similarUnverifiedMiniMax H3 has been integrated into mainstream AI workflow tools such as ComfyUI.75% similarUnverifiedThe generation phase of this workflow relies on the Minimax H3 Max model on the Fal.ai platform73% similarPartially VerifiedMinimax has released a new model, M3, claiming top-three performance globally, with support for 1M context and native multimodal capabilities.72% similarUnverifiedMiniMax H3 is a multimodal generative model with two core inference modes: REF2VA and FL2VA.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/954298API
curl https://kongchang.com/api/v1/knowledge/claims/954298MCP
get_claim(id=954298)