Breaking

Fish Audio Opens Top Voice Model S2.1 Pro Free to Developers

July 9, 2026 at 14:45 EDT

Media Generation
Foundation Models

On June 23, 2026, Fish Audio announced it had released its most advanced text-to-speech (TTS) model, "S2.1 Pro," as a free API for developers. It supports 83 languages, has no hard usage cap, and delivers paid-tier quality at no additional cost.

June 23, 2026 · Fish Audio

Paid-Tier Voice AI, Now Free: S2.1 Pro Ships as a No-Cap API

Fish Audio releases its most advanced text-to-speech model as a free API — 83 languages, no hard usage cap, and the same quality as the paid tier. Switch by swapping a single model string.

languages supported

~90ms

Time-to-First-Audio (TTFA)

free through end of July 2026

Free tier price vs. the paid rate

Same top quality — the paid version is metered per million UTF-8 bytes.

s2.1-pro-free
(free)

$15

s2.1-pro (paid)
per 1M UTF-8 bytes

Free vs. Paid — what actually differs

s2.1-pro-free

Price: free (until end of July 2026)
Same quality as paid
No hard cap (Fair Use)
No SLA / latency guarantee

s2.1-pro (paid)

$15 / million UTF-8 bytes
Top quality
Metered usage
SLA + latency guarantee

Built on S2 Pro (~4B params, Dual-AR, 10M+ hrs)

61%

win rate vs. the prior generation

2×+

more than double the throughput

1 line

swap a model string to switch from existing integrations

Emotion & prosody via natural-language tags like [whispers sweetly] · multi-speaker dialogue · voice cloning. High-revenue use above $1M ARR requires separate contact.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.