async Flash v1.0
Listed Proprietary · Async
5.8
/100
async Flash v1.0 is a proprietary text-to-speech (TTS) API from Async, measured by one independent benchmark, Artificial Analysis.
Artificial Analysis
Hume AI Octave 2 Hume AI · text-to-speech (TTS) API
Listed448 ms P50 time-to-first-audio on Speko; ranked 69th (tied) of 112 in the TTS for Voice Agents 2026 index.
Benchmark summary: 5.8/100, ranked 69th (tied) of 112 in the TTS for Voice Agents 2026 index (Listed). Strongest criterion: voice quality, 3.9 of 20 points.
Hume AI Octave 2 is a proprietary text-to-speech (TTS) API from Hume AI, measured by two independent benchmarks: Artificial Analysis and Speko.
Overall score
5.8 /100
Score breakdown
- Real-time speed 0.6 /60
- Voice quality 3.9 /20
- Accuracy 0.8 /15
- Price 0.5 /5
Data checked on September 30, 2026
Updated after each benchmark capture
Every point comes from a public benchmark: sources · methodology
Hume AI Octave 2 ranks 69th of 112 in the 2026 text-to-speech (TTS) API benchmark, with a score of 5.8/100.
Real-time latency
Speko measures a P50 TTFA of 448 ms from US East (n=30), dated July 3, 2026: 23rd lowest of the 29 models Speko measures.
Accuracy (word error rate)
Speko rates its pronunciation robustness on numbers, dates and currency at 0.24 on a 0-to-1 scale (18th highest of the 28 models Speko rates).
Voice quality
In the Artificial Analysis Speech Arena (capture of October 1, 2026), Hume AI Octave 2 has an Elo of 1,048 (±12, 95% confidence) from 3,053 blind comparisons, 59th highest of the 92 models Artificial Analysis rates. Speko rates its naturalness at 1,377 Elo from blind A/B votes (40th highest of the 41 models Speko rates).
Price
Artificial Analysis lists Hume AI Octave 2 at $87.5 per 1M characters (67th lowest of the 78 models Artificial Analysis lists a price for). Speko lists a cost of about $100 per 1M characters.
Scoring note
Speko changed its pronunciation robustness between the captures of September 30, 2026 and October 1, 2026 without a new measurement date (0.2, then 0.24): the score uses the average, 0.22. The value shown is the latest.
Languages tested
Tested in Spanish, German, French and Japanese by Artificial Analysis or Speko. A language missing here was not tested, which does not mean it is unsupported.
Benchmark measurements
Speko
- P50 TTFA, US East
- 448 ms
- Naturalness (Elo)
- 1,377
- Head-to-head win rate
- 29%
- Pronunciation robustness (0 to 1)
- 0.24
- Voice drift (lower is steadier)
- 25
- Price per 1M characters
- 100 USD
- Measured on
- 2026-07-03
- Speko model
- octave-2
- Speko model page
- benchmarks.speko.ai
Artificial Analysis Speech Arena
- Speech Arena Elo (blind listener preference)
- 1,048
- Elo 95% interval (±)
- 12
- Arena rank
- 59
- Arena rank range
- 53-63
- Arena appearances
- 3,053
- Voices tested
- 8
- Price per 1M characters
- 87.5 USD
- Released
- Oct 2025
- Artificial Analysis model
- Octave 2
- Artificial Analysis leaderboard
- artificialanalysis.ai
Not collected
- Speed (Coval time to first audio) — not measured by Coval (capture of 2026-10-01, 30-day window); counts 0
- Accuracy (Coval word error rate) — not measured by Coval (capture of 2026-10-01, 30-day window); counts 0
Similar TTS models
Hume AI Octave TTS
Listed Proprietary · Hume AI
5.8
/100
Hume AI Octave TTS is a proprietary text-to-speech (TTS) API from Hume AI, measured by two independent benchmarks: Artificial Analysis and Speko.
Artificial Analysis Speko
StepFun Step Audio EditX (Mar 2026)
Listed Open weights · StepFun
6
/100
StepFun Step Audio EditX (Mar 2026) is an open-weights text-to-speech (TTS) API from StepFun, measured by one independent benchmark, Artificial Analysis.
Artificial Analysis