PulseAugur
EN
LIVE 23:40:10

MiniMax releases open-source H3 model for real-time video generation

MiniMax has released its H3 model, which is capable of generating video with synchronized audio in real-time. This model, built on vLLM-Omni and FastVideo's FastH3, can render a 10.1-second MP4 video in 8.7 seconds. The company emphasizes that this is an open-source release, intended to foster further development and improvement within the community, leveraging NVIDIA AI hardware. AI

IMPACT Accelerates real-time video generation capabilities and encourages community-driven improvements in open-source AI.

RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on X — MiniMax AI →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

MiniMax releases open-source H3 model for real-time video generation

How we ranked this

Signal score
19 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Significant
Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. X — MiniMax AI TIER_1 English(EN) · MiniMax_AI ·

    This is what open-source intelligence is for.

    This is what open-source intelligence is for. Built on vLLM-Omni and @haoailab's FastH3, with @NVIDIAAI hardware support - all public. Real-time generation makes interactive video possible; an open baseline makes it improvable for everyone. Take it and go faster! 🚀