MiniMax H3 Video Model Hits SGLang: Runs Native 2K Locally on Dual RTX 5090s
ying11231 · x · 2026-08-03
MiniMax's open video model H3 is now available on SGLang Diffusion with day-0 serving support. The model takes text, image, video, and audio inputs to generate 5-15 second clips at native 2K resolution, 24fps, with native stereo audio.
H3 matches the performance of Seedance 2.0 at one-third of the cost. For local deployment, it can run entirely free on just 2x RTX 5090 or 1x RTX 6000. SGLang Diffusion also supports NVIDIA Blackwell/Hopper architectures and AMD MI355X/MI300X.
Related event: MiniMax Video Model Goes Viral with Stunning User Tests(61 posts)→
More from Infra
- GitHub rewrote the 800k-line Copilot runtime in Rust using AI agents, cutting CPU and memory by over 90% — bibryam · 2026-10-16
- NVIDIA quietly delists the entry-level RTX 5050 from its marketplace, following the RTX 5090 — chemist_slime · 2026-10-16
- Grok claims 816B MoE model runs locally on dual M5 Ultra Mac Studios via MLX — grok · 2026-10-16
- Same 1,000 Cloudflare Durable Objects: $4,050/month cut to $22 with one trick — threepointone · 2026-10-16
- NVIDIA ships FP4-quantized Qwen3.8-Flash for local use, trending on Hugging Face — nvidia · 2026-10-16
- Baidu's Virtual Head Attention tames the KV cache, FLOPs, quality triangle in long-context LLMs — baidu · 2026-10-16