Fish Audio Makes S2.1 Pro Voice Model Free, Costs 1/6 of ElevenLabs
rohanpaul_ai · x · 2026-07-30
Fish Audio has made its S2.1 Pro voice cloning model free for a month. The model requires only 10-15 seconds of audio to clone a voice, supports 83 languages, and features a low latency of 90ms with word-level pronunciation and pause control.
Architecture & Ecosystem: It uses a Dual-AR design (a 4B-parameter model for semantics/emotion and a 400M-parameter model for acoustic details) balancing speed and expressiveness. Built on their open-source project Fish Speech, they continue to offer open-weight models for self-hosting. Pricing is roughly 1/6th of ElevenLabs.
Core Advantage: Tuned for real conversation, the model is designed to survive interruptions, corrections, laughter, and language switches.
Related event: Fish Audio Launches S2.1 Pro Voice Model with $52M Seed Funding(11 posts)→
More from Models
- NVIDIA open-weight model ties pplx-decider at #1 on 669-case clinical benchmark — MaziyarPanahi · 2026-10-16
- Claude Code adds idle compaction to save 1h prompt cache on idle sessions — EricBuess · 2026-10-16
- Thrive Holdings launches real-work AI evals from 75+ businesses, benchmarking 13 models — samaysham · 2026-10-16
- Microsoft's MAI-Transcribe now handles 61% of all transcription on OpenRouter — mustafasuleyman · 2026-10-16
- Reflection AI Releases Open-Weight 501B MoE Beam for Agentic Coding on AWS — AWS ML Blog · 2026-10-16
- Dev RL-tunes Qwen 27B against slop with Meta's XAR method, lands top-4 on EQbench — Mysterious_Hearing14 · 2026-10-16