Wednesday, 23 September 2026 | التحديث اليومي نظرة ثاقبة للذكاء الاصطناعي، مكتوبة للبناة

MiniMax H3 (Hailuo 3.0)

MiniMax H3 (Hailuo 3.0) — Specifications

Compiled by Mustafa Ihsan from the vendor’s published documentation · Last updated

المطوّر MiniMax
النوع توليد الفيديو
النمط Text, Image, Video, Audio -> Video
عدد المعايير (البارامترات) 33B
الترخيص MiniMax H3 Community License
أوزان مفتوحة المصدر نعم
أُطلِقَ 2026-07
مقدمو واجهات برمجة التطبيقات (API) MiniMax API, Hailuo AI app

الصفحة الرسمية →

ما هو MiniMax H3؟

MiniMax H3, branded Hailuo 3.0 in the Hailuo AI app, is MiniMax’s omni-modal video model, released on 31 July 2026. According to MiniMax, it understands text, images, video and audio together and generates clips of up to 15 seconds at up to 2K resolution with native stereo sound, with native multi-shot planning, accurate text and brand rendering, and video-to-video motion transfer.

Its weights are published on Hugging Face as MiniMaxAI/MiniMax-H3 (about 33 billion parameters) under the MiniMax H3 Community License. MiniMax claims its per-second price at 2K is under a third of mainstream models’; it has not published a per-second API rate on the announcement page, so it is not in the API cost calculator. A plain-language guide: MiniMax H3 (Hailuo 3.0) explained.

MiniMax H3 (Hailuo 3.0) pricing

MiniMax H3 (Hailuo 3.0) has no public per-token API price. It is open-weight, so the only cost is the hardware you run it on — see the local-hardware table above and size a GPU with the حاسبة الذاكرة VRAM.

انتقل إلى الأعلى