Skip to content

Alibaba (Qwen) · read on October 27, 2026

Alibaba Qwen3 TTS Fast

Alibaba Qwen3 TTS Fast is a TTS (text-to-speech) model from Alibaba (Qwen). Coval measures its P95 time to first audio at 98.8 ms, which puts it in the Headroom speed tier. It ranks 4 of 86 in the full ranking.

Speed tier
Headroom
P95 / median
98.8 ms / 62.9 ms
Listener Elo
Not rated
Price per 1M chars
$10

Where it ranks

Full ranking: 4 of 86. Use case ranks, by speed:

The tier and the ranks are computed by this site from the benchmarks' figures, not declared by Alibaba (Qwen).

Measurements and sources

Coval

Median time to first audio (P50)62.9 ms
Time to first audio, P9598.8 ms
Time to first audio, mean68.2 ms
Time to first audio, P9080.7 ms
Timed requests2,808

Speko

Naturalness score1,536
Latency, board chart37 ms
Latency, model page sweep (median)64 ms
Numbers and dates pronunciation0.3

Benchmarked through Nari Labs's API: the latency shown is the host's, not the model creator's.

Details

ProviderAlibaba (Qwen)
AccessOpen weights
StreamingA benchmark measures its time to first audio
Features listed by Covalmultilingual
Languages testedNot collected: no benchmark here lists languages for this model
Measured byCoval, Speko
Official siteqwen.ai

Not collected

Nothing below is estimated. Each line says which measure is missing and why.

Sources

Values are copied from these pages on the date shown. The match between this model and its benchmark entries is inferred, because a source writes the name differently: Speko: “Nari Qwen3-TTS Fast”. See the methodology.

Compare it with the reference model: Gradium TTS vs Alibaba Qwen3 TTS Fast.