Aller au contenu

TTS API comparison 2026

Gradium TTS vs Inworld Realtime TTS-2 Gradium vs Inworld · text-to-speech for voice agents

Gradium TTS is in the Headroom speed tier (P95 at or under 101 ms); Inworld Realtime TTS-2 is in the Fast enough speed tier (P95 from 138 ms to 263 ms). Below, each measure side by side with its source; a model leads on a measure only beyond the tie rule. No combined score, no overall winner.

Data read October 22, 2026 · Coval, Speko, Artificial Analysis

Side by side

Measure Gradium TTS Inworld Realtime TTS-2
Speed tierCoval P95, 30 days HeadroomFast enough
Median time to first audioCoval, 30-day window 55 ms 153 ms
P95 time to first audio (slow turns)Coval, 30-day window 101 ms 200 ms
Time to first audio, board chartSpeko, US East 55 ms 84 ms
Naturalness EloSpeko 1,592 1,561
Listener Elo (95% interval)Artificial Analysis 1,176 ±13 1,267 ±15
Numbers and dates (0 to 1)Speko 0.27 0.61
List price per 1M charactersCoval, Speko, Artificial Analysis $35.9 to $57.8 $15 to $25

In bold: the model ahead on that measure. “tie”: latencies within 10 ms, or Elo scores whose 95% intervals overlap.

Where Gradium TTS is ahead

  • →Median time to first audio: 55 ms vs 153 ms (Coval, 30-day window)
  • →P95 time to first audio (slow turns): 101 ms vs 200 ms (Coval, 30-day window)
  • →Time to first audio, board chart: 55 ms vs 84 ms (Speko, US East)
  • →Naturalness Elo: 1,592 vs 1,561 (Speko)
Gradium TTS: all measurements →

Where Inworld Realtime TTS-2 is ahead

  • →Listener Elo (95% interval): 1,267 ±15 vs 1,176 ±13 (Artificial Analysis)
  • →Numbers and dates (0 to 1): 0.61 vs 0.27 (Speko)
  • →List price per 1M characters: $15 to $25 vs $35.9 to $57.8 (Coval, Speko, Artificial Analysis)
Inworld Realtime TTS-2: all measurements →

Frequently asked questions

What is the difference between Gradium TTS and Inworld Realtime TTS-2?
Gradium TTS (Gradium) is in the Headroom speed tier (P95 at or under 101 ms); Inworld Realtime TTS-2 (Inworld) is in the Fast enough speed tier (P95 from 138 ms to 263 ms). Gradium TTS leads on median time to first audio, P95 time to first audio (slow turns), time to first audio, board chart and naturalness Elo; Inworld Realtime TTS-2 leads on listener Elo (95% interval), numbers and dates (0 to 1) and list price per 1M characters. No combined score is computed.
Which is faster for real-time voice agents, Gradium TTS or Inworld Realtime TTS-2?
On Coval's 30-day window read on October 22, 2026, the median time to first audio is 55 ms for Gradium TTS and 153 ms for Inworld Realtime TTS-2 (Gradium TTS ahead by 98 ms); the P95, the slow turn, is 101 ms and 200 ms (Gradium TTS ahead by 99 ms). Gaps under 10 ms are ties.
Which sounds more natural, Gradium TTS or Inworld Realtime TTS-2?
Speko naturalness: 1,592 for Gradium TTS, 1,561 for Inworld Realtime TTS-2. Artificial Analysis listener Elo: 1,176 ±13 and 1,267 ±15. The two sources use different listeners and can disagree.
How are these two text-to-speech APIs compared?
Speed first, on Coval's median and P95 time to first audio; then voice quality (Speko, Artificial Analysis), accuracy on numbers and dates (Speko) and list price. Latencies within 10 ms and Elo scores whose intervals overlap are ties. Scores from different benchmarks are never added into one number.
Can these numbers be checked?
Yes. Each model page lists every measurement with its source and the date of the reading, and links to the benchmark page it comes from. The rules are on the methodology page.

More comparisons

Compare two other TTS APIs →