Vibeleaderboard
← All Intel
Intel / post

HeyGen Voice tops the controlled-voice TTS arena

Source
x.com
Date
ArtificialAnlys@ArtificialAnlys

HeyGen Voice takes the #1 spot on the Artificial Analysis Controlled Voice TTS Arena Leaderboard, ahead of Alibaba's Qwen-Audio-3.1-TTS-Plus and ElevenLabs' Eleven v4 Turbo HeyGen Voice is the first Text to Speech model from @HeyGen evaluated in our Controlled Voice Arena, where models are evaluated using the same 8 cloned voices across US and UK English. Key takeaways: ➤ Controlled Voice: HeyGen Voice has an Elo of 1,201 across 1,468 arena appearances, placing it ahead of Qwen-Audio-3.1-TTS-Plus at 1,182, Eleven v4 Turbo at 1,166 and Eleven v4 at 1,162. It ranks #1 in Assistants and Customer Service, #2 in Knowledge Sharing and #1 in English (UK). ➤ Pronunciation Robustness: HeyGen Voice scores 83.1%, ranking #10 of 29 models, behind Qwen-Audio-3.1-TTS-Plus at 84.6% and ahead of Inworld Realtime TTS-2 at 80.7%. It ranks #2 in Preserving Exact Sequences at 83.3%, behind only SpaceXAI TTS. ➤ Pricing: HeyGen Voice is priced at $30/1M characters, compared to Qwen-Audio-3.1-TTS-Plus at $19.3/1M characters, Eleven v4 Turbo at $40/1M characters and Eleven v4 at $80/1M characters. ➤ Speed: HeyGen Voice processes 40 characters per second of generation time, compared to 76 characters…

Read the full post on X
Why it matters

For voice builders, it gives comparable numbers: HeyGen Voice leads at Elo 1,201 for $30 per 1M characters, versus Qwen at $19.3 and ElevenLabs v4 Turbo at $40, but generates slower at 40 characters per second.

Key takeaways · AI-distilled
  • The Controlled Voice arena scores every model on the same 8 cloned voices across US and UK English; HeyGen Voice's Elo of 1,201 comes from 1,468 arena appearances, ahead of Eleven v4 Turbo (1,166) and Eleven v4 (1,162).
  • By use case, Artificial Analysis ranks HeyGen Voice #1 in Assistants and Customer Service, #2 in Knowledge Sharing, and #1 for UK English.
  • Pronunciation robustness is mid-pack: 83.1%, #10 of 29 models, behind Qwen-Audio-3.1-TTS-Plus at 84.6%, though it ranks #2 at preserving exact sequences (83.3%), behind only SpaceXAI TTS.
  • Its 40 characters per second of generation trails Eleven v4 at 76 and Gemini 3.8 Flash TTS at 46, so the arena lead comes with slower output.
Terms in this piece · Glossary
  • AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
More from ArtificialAnlys
Recommended reads
Comments

Checking sign-in…

Loading comments…