Speech Arena Elo 1,130 on Artificial Analysis, latency not measured; ranked 54th of 112 in the TTS for Voice Agents 2026 index.
Benchmark summary: 8.3/100, ranked 54th of 112 in the TTS for Voice Agents 2026 index (Listed). Strongest criterion: voice quality, 7.4 of 20 points.
StepFun Step TTS 2 (Mar 2026) is a proprietary text-to-speech (TTS) API from StepFun, measured by one independent benchmark, Artificial Analysis.
Overall score
8.3 /100
Data checked on September 30, 2026
Updated after each benchmark capture
Every point comes from a public benchmark: sources · methodology
StepFun Step TTS 2 (Mar 2026) ranks 54th of 112 in the 2026 text-to-speech (TTS) API benchmark, with a score of 8.3/100.
Neither Coval nor Speko measures StepFun Step TTS 2 (Mar 2026)'s latency, so it scores 0 of 60 on real-time speed and does not appear in the real-time ranking.
No accuracy measurement: Coval does not measure its word error rate and Speko does not rate its pronunciation robustness, so accuracy counts 0 of 15.
In the Artificial Analysis Speech Arena (capture of October 1, 2026), StepFun Step TTS 2 (Mar 2026) has an Elo of 1,130 (±13, 95% confidence) from 2,310 blind comparisons, 25th highest of the 92 models Artificial Analysis rates.
Artificial Analysis lists StepFun Step TTS 2 (Mar 2026) at $40 per 1M characters (49th lowest of the 78 models Artificial Analysis lists a price for).
Artificial Analysis changed its Speech Arena Elo between the captures of September 30, 2026 and October 1, 2026 without a new measurement date (1,134, then 1,130): the score uses the average, 1,132. The value shown is the latest.
async Pro v1.0 is a proprietary text-to-speech (TTS) API from Async, measured by one independent benchmark, Artificial Analysis.
Rime arcanav3 is a text-to-speech (TTS) API from Rime, measured by one independent benchmark, Speko.
ElevenLabs Eleven v3 is a proprietary text-to-speech (TTS) API from ElevenLabs, measured by one independent benchmark, Artificial Analysis.