Inworld Realtime TTS-2 Flash Inworld · text-to-speech (TTS) API

Speed

On Coval's 30-day window ending October 22, 2026 (3,180 runs from US East), Inworld Realtime TTS-2 Flash starts speaking after 61 ms on a typical turn (median) and 98.2 ms on a slow turn (P95, the slowest one in twenty). That P95 is 3rd lowest of the 32 models Coval measures, which puts it in the Headroom tier: headroom to spend elsewhere in the turn. On the median it is a tie (gap under 10 ms) with Alibaba Qwen3 TTS Fast and Gradium TTS. Speko's board plots it at 50 ms time to first audio from US East; its model page reports a 82 ms sweep.

Voice quality

Speko rates its naturalness at 1,501 Elo from blind A/B votes on phone-agent lines (24th highest of the 43 models Speko rates). In the Artificial Analysis Speech Arena it holds 1,218 ±13 Elo (95% interval), 12th highest of the 98 models Artificial Analysis rates; scores whose intervals overlap are a tie. In its Customer Service category, Artificial Analysis places it 19th, at 1,236 Elo.

Numbers and dates

Speko's pronunciation test on numbers, dates and currency scores it 0.58 on a 0-to-1 scale (7th highest of the 30 models Speko scores). This is the accuracy measure used here; Coval's word error rate separates models too little to compare them.

Price

Coval lists $15 per 1M characters (since Aug 24, 2026). Artificial Analysis lists $10.4. Speko lists about $15. List prices, as each source prints them: a range of $10.4 to $15.

Languages tested

Tested in Spanish, German, French, Portuguese, Japanese, Chinese, Arabic, Hindi and Vietnamese by Artificial Analysis or Speko. A language missing here was not tested, which does not mean it is unsupported.