PulseAugur
EN
LIVE 02:53:34
中文(ZH) MiniMax H3正式发布

MiniMax releases H3, a multimodal generative model with audio-visual output

MiniMax has officially launched its new general-purpose, multimodal generative model, MiniMax H3. This model is capable of understanding unified multimodal contexts including text, images, video, and sound, and can output native dual-channel audio-visual content up to 15 seconds at 2K resolution. MiniMax plans to release the model weights in the coming days, adhering to legal regulations. AI

IMPACT This new multimodal model could advance AI's ability to process and generate complex, integrated media formats.

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

Read on 36氪 (36Kr) →

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

MiniMax releases H3, a multimodal generative model with audio-visual output

COVERAGE [1]

  1. 36氪 (36Kr) TIER_1 中文(ZH) ·

    MiniMax H3 Officially Released

    36氪获悉,MiniMax宣布正式发布MiniMax H3。据介绍,这是一款通用的全模态生成模型,支持对文本、图像、视频、声音组成的多模态上下文的统一理解能力、能够输出具备原生双声道的音视频,最高可支持15s 2K分辨率。MiniMax表示,计划在未来几天内,在符合相关法律法规的前提下,开放模型权重。