FLUX 3 Max 在 AA-Video-T2V v2.1 物理能力登顶
FLUX 3 Max 在 Artificial Analysis 的 AA-Video-T2V v2.1 评测中，Physics 得分为所有模型最高，Audio Synchronization 和 Dialogue & Lip Sync 最接近前沿水平，而 Camera Control 和 Human Anatomy 差距最大。
Where FLUX 3 Max sits against the frontier on AA-Video-T2V v2.1: highest in Physics, closest behind in Audio Synchronization and Dialogue & Lip Sync
AA-Video-T2V v2.1 measures 10 capabilities and 10 use cases, each with its own leaderboard. Capabilities are the model skills a prompt is written to test.
FLUX 3 Max scores highest of any model in Physics, and sits closest behind the frontier in Audio Synchronization and Dialogue & Lip Sync. It sits furthest from it in Camera Control and Human Anatomy. Against its base model FLUX 3, it scores higher on seven of the ten capabilities, by the most in Lighting & Materials and Complex Composition.
➤ Physics covers mechanics, thermal effects, material interactions, fracture and deformation, optics, cause and effect, and counterfactual physics.
➤ Dialogue & Lip Sync covers single- and multi-speaker dialogue, and voiceover vs on-screen speech.
➤ Audio Synchronization covers diegetic sound effects, background music and score, and on- and off-screen sound.
来源：ArtificialAnlys · x.com