Seedance 2.0 and HappyHorse-1.0 occupy top two Artificial Analysis video leaderboard slots — Veo 3.1 drops to #3, structural Chinese-vendor leadership in video-generation
ByteDance's Seedance 2.0 (released February) and Alibaba ATH's HappyHorse-1.0 (released April) occupy the top two positions on the Artificial Analysis text-to-video leaderboard, displacing Veo 3.1 to #3. The H1 2026 structural pattern in video generation: Chinese vendors lead, US/European frontier labs (Google Veo, Runway, Pika) follow.
The substantive piece is the structural-leadership shift in the video-generation category specifically. Through 2024-2025 Google Veo, Runway, and Pika collectively defined the video-generation frontier; Chinese vendor releases (Kling, MinMaxv1) competed at second-tier positioning. The Seedance 2.0 + HappyHorse-1.0 leaderboard dominance through H1 2026 reverses that pattern. The cause appears to be a combination of training-data-scale advantage and unified-modality-architecture decisions (Seedance 2.0's audio-visual-sync-at-generation, HappyHorse's multi-shot capability).
The procurement read for video-generation workloads is that vendor selection should now weight Chinese-vendor offerings as the leading-tier option rather than the competitive-tier option. Runway's pivot toward video editing reads partially as a competitive response — if Runway can't lead in pure generation, the editing-specialist position is a defensible niche.
Pinggy — Best Video Generation AI Models in 2026 → · Pixflow — Best AI Video Generators in 2026 - Free & Paid Ranked → · Pixazo — Best AI Video Generation Models in 2026: In-Depth Comparison →