// news · multimodal2026-06-16source: kling / artificial analysis / llm-stats

Kling v3 leads text-to-video arena leaderboard with 2031 Elo as Runway Gen-4.5 drops out of top 10 — Chinese-physics-quality premium stabilizes through mid-June

Kling v3 leads the text-to-video arena leaderboard with an Elo of 2031, followed by LTX-2 Fast (1930) and Happy Horse 1.0 (1885). Veo 3.1 holds at #3 with audio; Kling 3.0 has four entries in the top 10. Runway Gen-4.5 has dropped out of the top 10. The Chinese-physics-quality premium that emerged in Q1 2026 is structurally durable through mid-June.

The substantive piece is the quality-tier durability. Blind-vote arena leaderboards remove brand-driven evaluation bias — users score outputs without seeing which model produced them. Kling v3's sustained Elo lead through mid-June after ByteDance Seedance 2.0 (Feb) and Alibaba HappyHorse-1.0 (April) entered the field demonstrates that the Chinese video-generation quality premium is genuine rather than benchmark-gamed. Veo 3.1 holds the audio-synchronization niche; Runway Gen-4.5's drop out of the top 10 reflects the workflow-vs-capability tradeoff playing out at scale.

The structural read against Runway Gen-4.5's marketer-workflow specialization pattern is that the video-generation procurement segmentation is now clean: capability buyers route to Kling 3.0 / Veo 3.1 / LTX-2 by use case, workflow buyers route to Runway by integration pattern. The 2024-2025 single-leader consolidation narrative is structurally dead through 2027.

See our analysis →

LLM Stats — Best AI for Video Generation in 2026 — Ranked by Blind Human Votes → · Pinggy — Best Video Generation AI Models in 2026 → · Atlas Cloud — Best AI Video Generation Models in 2026: Complete Comparison →