CoeFont Releases v4 Voice Model With Faster Generation
CoeFont is selling speech velocity rather than a new voice set, which points to generation time, not fidelity, as the capability vendors on this beat now compete on.
Reporting from 1 source: ASCII.jp.
CoeFont, Inc. released CoeFont v4 on September 11, 2026, a new voice model for its AI voice platform. The company reports generation is up to 61 times faster than the previous version, with time cut by as much as 98 percent. Internal measurements list 0.08 seconds for short text, 0.10 seconds for medium text, and 0.17 seconds for long text.
CoeFont v4 replaces the model that runs the platform's voice generation, and the company's own numbers put the gains in time rather than accuracy: short text drops from 3.62 seconds to 0.08 seconds, medium from 4.62 to 0.10, and long from 10.36 to 0.17. CoeFont, Inc. is a Tokyo Institute of Science-certified venture established in 2020. The platform's Voice Hub lists more than 10,000 AI voices, and the company says CoeFont is available in all countries and regions.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.