The open-source video generation stack just gained a second layer. Weeks after MiniMax released the H3-Base weights, the FastVideo team at UCSD, working with NVIDIA's FastGen team and Nuva Lab, shipped FastH3 Preview v1 — a family of open-weight checkpoints post-trained on MiniMax H3 that push text-to-audio-video (T2VA) generation into a new latency class.
From 49 transformer calls down to 4
The cost problem is stark. Base H3's audio-video diffusion transformer has 33B parameters, and generating one video invokes it 49 times. FastH3 attacks both the number of calls and the work inside each call.
The first cut is DMD2 (Distribution Matching Distillation). A frozen Base H3 teacher and a learned critic train the student; the difference between their score estimates supplies the training signal, and backward simulation exposes prompt-only students to the few-step states they will see at inference. The result: five scheduler points, exactly four DiT forwards at generation time.
The second cut is VSA (Video Sparse Attention). The trainable sparse-attention student keeps about 10% of eligible video-to-video tiles (90% sparsity, 64-token blocks) while text and audio stay dense; the teacher and critic also run dense attention, giving the sparse student a full-attention target.
The numbers: up to 14.38x on a single B200
At 1344×768, 24 FPS with audio, Base H3 with dense FA4 takes 678.7 seconds for a 15-second clip on one B200; the recommended VSA/Data-Free checkpoint takes 47.2 seconds — a 14.38x speedup. A 5-second clip: 132.5s vs 16.2s, 8.16x. On 4× B200 a 5-second clip lands at 6.1 seconds end-to-end; on 8× B200 a 15-second clip lands at 12.88 seconds — faster than the clip itself plays. Timings are the median of three requests after warmup, with model load and compilation excluded, and end-to-end includes encoding, denoising, decoding, audio, muxing, and file output.
Training consumed 1k+ B200 hours. The recommended checkpoint was trained to step 1300 and ships both as full weights and as a pre-extracted rank-64 LoRA.
Data-free training, stated limits
The recommended checkpoint is data-free: trained from prompts only, without target videos. That sidesteps video-data sourcing and licensing headaches entirely. The team also publishes synthetic-data ablations and a dense-attention reference — four checkpoints in the Preview v1 family, with training code promised soon.
The limits are stated plainly: T2VA only (FL2VA and Ref2VA are not distilled yet), difficult motion, fine detail, and some audio may fall below Base H3 quality, and everything inherits the MiniMax H3 Community License.
Why this matters beyond "another speedup"
It validates the two-layer innovation model that open weights enable: MiniMax shipped the base, and the community now inspects it, post-trains it, swaps kernels, and runs it on its own hardware. fal's H3 Max, released the same week, took the hosted-API route — the two paths now have a direct comparison.
Distillation plus sparse attention is becoming the standard route to production video generation. 49 calls down to 4, plus 90% attention sparsity, means the same hardware serves far more requests — or the same task fits into much smaller machines. Next on the roadmap: an 8-step quality variant, FP8/NVFP4 quantization, local optimizations for RTX, DGX Spark, and Apple MLX, and a new Parallel Decoding Distillation (PDD) collaboration with NVIDIA.
If you are evaluating open video stacks, run the recommended VSA/Data-Free checkpoint: the weights are hardware-independent, the 4× B200 setup is a controlled benchmark rather than a requirement, and the GPU count just needs to divide H3's 56 attention heads.
When generation time drops below the video's own duration, interactive and real-time workflows stop being demos and start being schedulable work. That is the real weight of this release.