#Video

AIArtificial Analysis reports that Grok Imagine Video 1.5 Lite comes closest to the frontier on AA-Video-T2V v2.0 in Multi-Scene & Narrative, Lighting & Materials, and Text Rendering. It is furthest behind in Dialogue & Lip Sync and Human Anatomy. Compared with Grok Imagine Video 1.5, Lite matches it in Physics and trails on the other nine capabilities, by the least in Multi-Scene & Narrative.

AIStability AI's Interactive Research team introduced SemanTok, which makes early video tokens more semantically meaningful so the representation is easier to predict. According to the post, a model using SemanTok matches or beats the performance of a model more than three times its size. The approach targets more efficient autoregressive video generation.

AIGoogle Developers Blog describes how Sparse VideoGen (SVG) routes video diffusion attention heads into spatial or temporal sparse masks and implements them as custom JAX and Pallas Splash Attention kernels on TPU v6e. In isolated single-chip tests with 75.6K tokens and 10 heads, the sparse variants retain about 38.87% of query-key pairs. The article argues that theoretical sparsity must be converted into hardware tile skipping to yield real speedups.

AIHiggsfield AI released "Passport Rush," described as the first hybrid AI-animated short made with a Blender and Higgsfield pipeline. The film combines handcrafted 2D storyboards and 3D Blender previs with AI animation. The project is now open source, with every prompt and asset publicly available.

AIAmazon Neuron Science and Reactor collaborated to optimize Rolling Forcing, an autoregressive diffusion video model, for real-time generation on AWS Trainium. The model generates 16 frames per second, the standard video playback rate, meeting latency requirements for streaming video generation.

AIGoogle Research announced a new unified multi-agent framework that helps creators generate temporally consistent long-form video narratives. It is designed to reduce visual drift and the propagation of errors across the production pipeline.

AIMeta's Muse Realtime Voice generates speech tokens encoding both content and prosody, and Muse Realtime Avatar consumes that shared stream to produce streaming video. Using a fixed-length history as motion context keeps computation bounded regardless of conversation length while synchronizing voice, lip motion, and expressions.

AIMeta distilled a 40-step diffusion teacher using 3-way CFG, requiring 120 evaluations per video chunk, into an unguided 2-step causal student with a fixed-length KV cache. The student uses self-forcing to resist drift and keep near-teacher quality while needing 60x fewer evaluations, enabling instant responses in live video streaming.

AIMiniMax-H3 on 8×H200 GPUs reaches 1.85–1.95x lossless speedup over Diffusers without approximation, with fixed prompts, seeds, resolution, FPS, and 50 denoising steps. Adding step reuse and sparse attention raises speedup to as much as 6.24x, but quality varies by workload, with SSIM from 0.76 to 0.91. Two presets trade off the two: a conservative Cache-DiT setting gives 2.99x at 0.90–0.98 SSIM, while a faster SubBlock 0.75 plus Cache-DiT stride gives 4.90–5.93x at 0.77–0.92.