ModelsAug 27, 2026 · 06:506 sources

fal H3 Max Takes #1 Image to Video Rank, 50x Faster Than MiniMax H3

From huggingnews.com

Image: huggingnews.com
AI;DR

fal Research released H3 Max, a post-trained model that generates 5 second 720p video clips in under 3 seconds for $0.04 per second of 768p resolution, ranking first in Image to Video and third in Text to Video on Artificial Analysis leaderboards. The model, built on the open weight MiniMax H3, is nearly 50x faster than the base H3 model and fal Research intends to release its weights.

Read the sourceSign in to suggest

Sources

+ 5 more
  1. SUPPORT@fal“roughly 35x the throughput of the official H3 endpoint”x.com
  2. SUPPORT@fal“tuning specifically for stronger prompt adherence and aesthetics”x.com
  3. SUPPORT@fal“designed the architecture of H3 Max around fal’s in-house inference engine”x.com
  4. SUPPORT@artificialanlys“$0.04 per second of 768p video ($2.40 per minute)”x.com
  5. SUPPORT@designarena“Image-to-Video generation time is just 6.4 seconds, 18x faster than average”x.com
#video#inferenceMarkdown
Content history (2)
  • 2026-10-16 · Summary · vi · Replaced stored text
    Mô hình H3 Max sau khi được post-trained có khả năng tạo các clip video 720p dài 5 giây t…
    Mô hình H3 Max đã được post-trained, tạo video 720p 5 giây trong dưới 3 giây với giá $0.0…
  • 2026-10-16 · Title · vi · Replaced stored text
    fal H3 Max chiếm ngôi đầu về chuyển ảnh thành video, tốc độ nhanh gấp 50 lần MiniMax H3
    fal H3 Max chiếm ngôi #1 Image to Video, nhanh 50x so với MiniMax H3

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today