VISTA: A Test-Time Self-Improving Video Generation Agent

Despite rapid advances in text-to-video synthesis, generated video quality remains critically dependent on precise user prompts. Existing test-time optimization methods, successful in other domains, struggle with the multi-faceted nature of video. In this work, we introduce VISTA, a novel multi-agent system that autonomously improves video generation through refining prompts in an iterative loop. VISTA first decomposes a user's idea into a structured temporal plan. After generation, the best video is identified through a robust pairwise tournament. This winning video is then critiqued by a trio of specialized agents focusing on visual, audio, and contextual fidelity. Finally, a reasoning agent synthesizes this feedback to introspectively rewrite and enhance the prompt for the next generation cycle. Experiments on single-and multi-scene video generation scenarios show that while prior methods yield inconsistent gains, VISTA consistently improves video quality and alignment with user intent, achieving up to 60% pairwise win rate against state-of-the-art baselines. Human evaluators concur, preferring VISTA's outputs in 66.4% of comparisons. Direct Prompting (DP) VISTA (Ours) Single-scene (Polyak et al., 2025): The person's forehead creased with worry as he listened to bad news. Single-scene (Polyak et al., 2025): A spaceship entering hyperdrive, stars streaking past as it accelerates. Multi-scene-Interviews: The video features a man outdoors, asking a trivia question about a comedian known. . . Multi-scene-Animation: The video is an educational animation designed for young children. . .

Table 1 | Example videos generated by VISTA, showing improvements in visual fidelity, camera focus, smooth transitions, compelling storylines, and sounds. Optimized prompts and more examples are in our project page.