Artificial Curiosity Labs

Speech models

Speech recognition, text-to-speech, and speech-to-speech quality.

Last validated snapshot: October 27, 2026

Speech-to-text

Artificial Analysis WER Index · Lower is better

Eligible
61
Candidates
77
Coverage
79.2%
Rank Model Creator Artificial Analysis WER Index Since prior snapshot
1 Fun-Realtime-ASR-preview Alibaba 0.017 Not comparable
2 MAI-Transcribe-2 Microsoft AI 0.02 Not comparable
3 Voxtral Small Transcribe 3 Mistral 0.021 Not comparable

Text-to-speech

Text-to-speech Elo · Higher is better

Eligible
98
Candidates
98
Coverage
100.0%
Rank Model Creator Text-to-speech Elo Since prior snapshot
1 Eleven v4 ElevenLabs 1,334 Not comparable
2 Eleven v4 Turbo ElevenLabs 1,333 Not comparable
3 Qwen-Audio-3.1-TTS-Plus Alibaba 1,306 Not comparable

Speech-to-speech BBA

BBA score · Higher is better

Eligible
42
Candidates
47
Coverage
89.4%
Rank Model Creator BBA score Since prior snapshot
1 StepAudio 3 Realtime StepFun 1 No change
2 tie Qwen3.5 Omni Plus Realtime Alibaba 0.99 No change
2 tie Eleven Realtime v1 Max ElevenLabs 0.99 Not comparable
2 tie Qwen Audio 3.0 Realtime Plus Alibaba 0.99 No change

Speech-to-speech FDB

FDB score · Higher is better

Eligible
34
Candidates
47
Coverage
72.3%
Rank Model Creator FDB score Since prior snapshot
1 StepAudio 3 Realtime StepFun 0.99 No change
2 Qwen Audio 3.0 Realtime Plus Alibaba 0.98 No change
3 tie GPT-Live-1 (Sol, low) OpenAI 0.97 No change
3 tie Grok Voice 3.0 SpaceXAI 0.97 Not comparable
3 tie Qwen Audio 3.0 Realtime Flash Alibaba 0.97 No change

Speech-to-speech Tau Voice

Tau Voice score · Higher is better

Eligible
28
Candidates
47
Coverage
59.6%
Rank Model Creator Tau Voice score Since prior snapshot
1 GPT-Live-1 (Astra, medium) OpenAI 0.74 Improved 1
2 Grok Voice 3.0 SpaceXAI 0.72 Not comparable
3 Gemini 3.8 Live Extended Thinking (High) Google 0.69 Down 2

Data source: Artificial Analysis .