Artificial Analysis: Qwen-Audio-3.1-TTS-Plus leads US English text-to-speech, HeyGen Voice leads UK English and overall

Artificial Analysis says that in its Controlled Voice Arena, which scores every text-to-speech model on the same eight cloned voices (four US, four UK), Qwen-Audio-3.1-TTS-Plus leads the US English ranking, while HeyGen Voice does better in UK English and leads the overall leaderboard.

It reports that 12 of the 49 models show statistically significant differences between US and UK English (non-overlapping confidence intervals): seven do better in US English and five in UK English. The average gap across all models is 7 Elo points, which it reads as no strong overall accent bias. The largest gaps it lists:

• Lightning V3.1 Pro: 94 Elo points higher in US English
• Qwen-Audio-3.1-TTS-Plus: 66 points higher in US English
• Soniox TTS Real-Time v2: 51 points higher in UK English
• Speech 2.8 Turbo: 49 points higher in UK English

Artificial Analysis advises evaluating a speech model in the accent your use case needs rather than relying on the overall ranking. The rankings are its own.