Bland Speech v3
Listed Proprietary · Bland AI
13.4
/100
Bland Speech v3 is a proprietary text-to-speech (TTS) API from Bland AI, measured by two independent benchmarks: Artificial Analysis and Speko.
Artificial Analysis Speko
Smallest.ai Lightning v3.1 Smallest.ai · text-to-speech (TTS) API
Listed173 ms P50 time-to-first-audio on Speko; ranked 36th of 112 in the TTS for Voice Agents 2026 index.
Benchmark summary: 14/100, ranked 36th of 112 in the TTS for Voice Agents 2026 index (Listed). Strongest criterion: price, 3 of 5 points.
Smallest.ai Lightning v3.1 is a proprietary text-to-speech (TTS) API from Smallest.ai, measured by two independent benchmarks: Artificial Analysis and Speko.
Overall score
14 /100
Score breakdown
- Real-time speed 1.6 /60
- Voice quality 7.7 /20
- Accuracy 1.7 /15
- Price 3 /5
Data checked on September 30, 2026
Updated after each benchmark capture
Every point comes from a public benchmark: sources · methodology
Smallest.ai Lightning v3.1 ranks 36th of 112 in the 2026 text-to-speech (TTS) API benchmark, with a score of 14/100.
Real-time latency
Speko measures a P50 TTFA of 173 ms from US East (n=30), dated July 3, 2026: 13th lowest of the 29 models Speko measures.
Accuracy (word error rate)
Speko rates its pronunciation robustness on numbers, dates and currency at 0.25 on a 0-to-1 scale (17th highest of the 28 models Speko rates).
Voice quality
In the Artificial Analysis Speech Arena (capture of October 1, 2026), Smallest.ai Lightning v3.1 has an Elo of 1,021 (±13, 95% confidence) from 2,145 blind comparisons, 72nd highest of the 92 models Artificial Analysis rates. Speko rates its naturalness at 1,544 Elo from blind A/B votes (19th highest of the 41 models Speko rates).
Price
Artificial Analysis lists Smallest.ai Lightning v3.1 at $25 per 1M characters (41st lowest of the 78 models Artificial Analysis lists a price for). Speko lists a cost of about $17.5 per 1M characters.
Scoring note
Artificial Analysis changed its Speech Arena Elo between the captures of September 30, 2026 and October 1, 2026 without a new measurement date (1,022, then 1,021): the score uses the average, 1,021.5. The value shown is the latest.
Languages tested
Tested in Spanish, German, French and Hindi 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
- 173 ms
- Naturalness (Elo)
- 1,544
- Head-to-head win rate
- 58%
- Pronunciation robustness (0 to 1)
- 0.25
- Voice drift (lower is steadier)
- 0
- Deterministic output
- ✓
- Price per 1M characters
- 17.5 USD
- Measured on
- 2026-07-03
- Speko model
- lightning_v3.1
- Speko model page
- benchmarks.speko.ai
Artificial Analysis Speech Arena
- Speech Arena Elo (blind listener preference)
- 1,021
- Elo 95% interval (±)
- 13
- Arena rank
- 72
- Arena rank range
- 65-73
- Arena appearances
- 2,145
- Voices tested
- 7
- Price per 1M characters
- 25 USD
- Released
- Mar 2026
- Artificial Analysis model
- Lightning v3.1
- 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
Deepgram Aura 2 En
Listed Proprietary · Deepgram
14.4
/100
Deepgram Aura 2 En is a proprietary text-to-speech (TTS) API from Deepgram, served by Cloudflare, measured by one independent benchmark, Coval.
Coval
OpenAI GPT-4o mini TTS
Listed Proprietary · OpenAI
14.4
/100
OpenAI GPT-4o mini TTS is a proprietary text-to-speech (TTS) API from OpenAI, measured by two independent benchmarks: Coval and Speko.
Spanish German French Japanese
+7