October 9, 2026
HeyGen Voice tops the Artificial Analysis voice arena, ahead of Eleven v4 Turbo
On Oct 9, 2026, HeyGen Voice scored 1,201 on the Elo rating (the same scale used in chess) and took first place in the Controlled Voice Arena. Qwen-Audio-3.1-TTS-Plus scored 1,182 and Eleven v4 Turbo 1,166.
Artificial Analysis
@artificialanlys
HeyGen Voice takes first place on the Artificial Analysis Controlled Voice TTS Arena leaderboard, ahead of Qwen-Audio-3.1-TTS-Plus from Alibaba and Eleven v4 Turbo from ElevenLabs. HeyGen Voice is the first TTS model from @HeyGen that we evaluated in our Controlled Voice Arena, where models are tested on the same 8 cloned voices in American and British English. Key takeaways: ➤ Controlled Voice: HeyGen Voice scored an Elo of 1,201 across 1,468 arena matchups, higher than Qwen-Audio-3.1-TTS-Plus (1,182), Eleven v4 Turbo (1,166) and Eleven v4 (1,162). It ranks first in Assistants and Customer Service, second in Knowledge Sharing, and first in English (UK). ➤ Pronunciation Robustness: HeyGen Voice scored 83.1% and ranks 10th out of 29 models. That is below Qwen-Audio-3.1-TTS-Plus (84.6%) and above Inworld Realtime TTS-2 (80.7%). In Preserving Exact Sequences (83.3%) it ranks second, behind only SpaceXAI TTS. ➤ Pricing: HeyGen Voice costs $30 per 1M characters. Qwen-Audio-3.1-TTS-Plus costs $19.3 per 1M characters, Eleven v4 Turbo $40, and Eleven v4 $80. ➤ Speed: HeyGen Voice generates 40 characters per second. Eleven v4 does 76 characters per second and Gemini 3.8 Flash TTS 46. Full results and audio samples below ⬇️
· 41.4K views
This is the first HeyGen TTS model that Artificial Analysis has evaluated in the Controlled Voice Arena. All models read the same eight cloned voices in American and British English, so the comparison reflects the model itself. HeyGen Voice ranks first in Assistants and Customer Service, and second in Knowledge Sharing.
Pronunciation. This test tells a different story: 83.1% and 10th place out of 29 models. Qwen-Audio-3.1-TTS-Plus holds 84.6%, while HeyGen Voice ranks second in exact sequence preservation (83.3%), behind only SpaceXAI TTS.
Price and speed. HeyGen charges $30 per million characters. Qwen-Audio-3.1-TTS-Plus costs $19.3, Eleven v4 Turbo $40, and Eleven v4 $80. Generation runs at 40 characters per second, almost half the speed of Eleven v4 at 76. Gemini 3.8 Flash TTS reaches 46 on this metric.