Sopro V2 Turbo 2610 Released: Cleaner Voices, Still 120M
10/04/2026 — 10/04, 01:11·1 sources·1 reports
Story overview
On October 4, 2026, a post on Reddit's LocalLLaMA forum introduced Sopro V2 Turbo 2610, described as an interim update to the voice cloning model that had been released the month before. According to the post, one of the main problems users ran into with the earlier version was roughness or break-up on some cloned voices. The 2610 build is focused mostly on fixing that, reducing roughness and break-up for the voices that struggled before.
The release keeps the same model size and speed: it is still a 120M-parameter model that reaches its first audio in roughly 300 ms on a laptop CPU. It ships under the Apache-2.0 license and supports English, European Portuguese, French, and German. The post notes that more languages and stronger generation controls are still planned, though no timeline is given.
As of this single report, that is where things stand: an interim update aimed at voice quality issues, with the parameter count, inference speed, and open-source license unchanged from the previous version.
AI-generated from 1 reports · updated 2 hours ago
Latest turnSopro V2 Turbo 2610 is an interim update to the voice-cloning model, aimed mainly at reducing roughness and break-up on cloned voices that struggled before. It keeps the same 120M model and roughly 300 ms to first audio on a laptop CPU, ships under Apache-2.0, and covers English, European Portuguese, French and German, with more languages and finer voice control planned.
Reports on this story headlines open the original
Sopro V2 Turbo 2610 is an interim update to the voice-cloning model, aimed mainly at reducing roughness and break-up on cloned voices that struggled before. It keeps the same 120M model and roughly 300 ms to first audio on a laptop CPU, ships under Apache-2.0, and covers English, European Portuguese, French and German, with more languages and finer voice control planned.
Reddit · LocalLLaMAAI score 72
Other stories people are talking about
- 527RisingApple tightens macOS Full Disk Access over AI agent risks9 sources
- 374NVIDIA launches 64GB DGX Spark desktop AI computer at $4,9998 sources
- 359RisingMeta open-sources Muse Gadgets firmware and SDK6 sources
- 276Claude Opus 5.5 and GPT-6 Sol: comparing cost per task4 sources
- 252Anthropic Reportedly Targets Nov 9 IPO Listing Before Thanksgiving5 sources
- 238Aleph Alpha releases Kolibri-1, a 78B MoE model with 1M context3 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion