naturalness benchmark
TTS Arena V2 is an official text-to-speech human preference leaderboard published by TTS-AGI.
Updated Sep 8, 2026
Higher bradley-terry elo ranks better on this benchmark.
Rank | Model | Bradley-Terry Elo | Percentile | Participants | Votes | Evidence | Evaluated |
|---|
| Rank01 | ModelIN | Bradley-Terry Elo1,557.00 elo | Percentile100.00% | Participants17 | Votes1,496 | EvidenceB | EvaluatedN/A |
| Rank02 | ModelIN | Bradley-Terry Elo1,539.00 elo | Percentile93.75% | Participants17 | Votes1,976 | EvidenceB | EvaluatedN/A |
| Rank03 | ModelSMLightning V3.1 Pro (Jul 2026)Smallest.ai | Bradley-Terry Elo1,536.00 elo | Percentile87.50% | Participants17 | Votes576 | EvidenceB | EvaluatedN/A |
| Rank04 | ModelMI | Bradley-Terry Elo1,528.00 elo | Percentile81.25% | Participants17 | Votes536 | EvidenceB | EvaluatedN/A |
| Rank05 | ModelHAOctave TTSHume AI | Bradley-Terry Elo1,523.00 elo | Percentile75.00% | Participants17 | Votes1,162 | EvidenceB | EvaluatedN/A |
| Rank06 | ModelMI | Bradley-Terry Elo1,521.00 elo | Percentile68.75% | Participants17 | Votes587 | EvidenceB | EvaluatedN/A |
| Rank07 | ModelMI | Bradley-Terry Elo1,520.00 elo | Percentile62.50% | Participants17 | Votes561 | EvidenceB | EvaluatedN/A |
| Rank08 | ModelMI | Bradley-Terry Elo1,510.00 elo | Percentile56.25% | Participants17 | Votes1,254 | EvidenceB | EvaluatedN/A |
| Rank09 | ModelMI | Bradley-Terry Elo1,508.00 elo | Percentile50.00% | Participants17 | Votes592 | EvidenceB | EvaluatedN/A |
| Rank10 | ModelFAOpenAudio S1Fish Audio | Bradley-Terry Elo1,508.00 elo | Percentile43.75% | Participants17 | Votes571 | EvidenceB | EvaluatedN/A |
| Rank11 | ModelEL | Bradley-Terry Elo1,505.00 elo | Percentile37.50% | Participants17 | Votes1,156 | EvidenceB | EvaluatedN/A |
| Rank12 | ModelEL | Bradley-Terry Elo1,505.00 elo | Percentile31.25% | Participants17 | Votes1,180 | EvidenceB | EvaluatedN/A |
| Rank13 | ModelEL | Bradley-Terry Elo1,505.00 elo | Percentile25.00% | Participants17 | Votes1,152 | EvidenceB | EvaluatedN/A |
| Rank14 | ModelMI | Bradley-Terry Elo1,503.00 elo | Percentile18.75% | Participants17 | Votes1,244 | EvidenceB | EvaluatedN/A |
| Rank15 | ModelHISpeech 2.6Hithink | Bradley-Terry Elo1,502.00 elo | Percentile12.50% | Participants17 | Votes585 | EvidenceB | EvaluatedN/A |
| Rank16 | ModelEL | Bradley-Terry Elo1,498.00 elo | Percentile6.25% | Participants17 | Votes587 | EvidenceB | EvaluatedN/A |
| Rank17 | ModelRAChatterboxResemble AI | Bradley-Terry Elo1,478.00 elo | Percentile0.00% | Participants17 | Votes1,793 | EvidenceB | EvaluatedN/A |
The leading models and scores on this benchmark.
A closer view of the leading scores on this benchmark.
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What TTS Arena V2 measures and how its scores work.
TTS Arena V2 is a text-to-speech human preference leaderboard.
It measures naturalness using Bradley-Terry Elo, reported in elo.
Scores are shown in elo. This benchmark is verified and has an evidence level of B.
TTS Arena V2. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
Common questions about TTS Arena V2.
Inworld TTS-1-Max is currently ranked first with 1,557.00 elo.
The current leaders are Inworld TTS-1-Max (1,557.00 elo), Inworld TTS-1 (1,539.00 elo), and Lightning V3.1 Pro (Jul 2026) (1,536.00 elo).
No matched official input price is currently available.
No matched runtime record is currently available.
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.
It measures naturalness using Bradley-Terry Elo, reported in elo.
Yes. Higher values rank better for this benchmark.
17 model results are currently shown.
No. This benchmark is shown for reference but does not contribute to the overall score.
Ranking basisThis tts arena v2 AI model leaderboard uses descending bradley-terry elo in the benchmark original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
Selection summary
Inworld TTS-1-Max currently leads TTS Arena V2 with 1,557.00 elo. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price, and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.