AI 导读
Gemini 3.1 Flash TTS 在 Contextually appropriate 上以 96.4% 领先，在 Expanding shorthand 上以 84.4% 领先，略高于 SpaceXAI TTS 的 84.3%。SpaceXAI TTS 在 Preserving exact sequences 上以 85.7% 领先，比 Realtime TTS-2 的 79.9% 高出 5.8 个百分点。Qwen-Audio-3.0-TTS-Plus 在 Standalone terms 上以 95.5% 领先。各模型之间差距最大的是 Preserving exact sequences，从 5.3% 到 85.7%。
正文
Gemini 3.1 Flash TTS leads Contextually appropriate at 96.4% and Expanding shorthand at 84.4%, just ahead of SpaceXAI TTS at 84.3%. SpaceXAI TTS leads Preserving exact sequences at 85.7%, 5.8 p.p. ahead of Realtime TTS-2 at 79.9%. Qwen-Audio-3.0-TTS-Plus leads Standalone terms at 95.5%. The widest spread between models is on Preserving exact sequences, from 5.3% to 85.7%.
来源：@ArtificialAnlys · x.com