Unverified50% confidenceReleaseExact time
The H3 speech generation model has released its full-precision weights.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/5/2026
First Seen
Valid until: 11/3/2026
Sources
H3语音模型全精度权重发布：表现力惊艳但长句声音漂移待解
Related Claims
UnverifiedH3 demonstrates strong voice cloning capability, with multiple users stating that its reproduction of specific individuals' voices and tones is very accurate.77% similarUnverifiedUser testing feedback indicates that H3's audio quality is outstanding and highly expressive.73% similarUnverifiedThe cloning capability demonstrated by H3 is zero-shot cloning, which extracts a speaker's acoustic characteristics from only a few seconds to tens of seconds of reference audio.66% similarUnverified该项目支持声纹识别和声音克隆功能61% similarUnverified声纹识别技术从GMM-UBM、i-vector发展到x-vector（基于TDNN）和d-vector（基于LSTM），噪声条件下识别准确率显著提升61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/687458API
curl https://kongchang.com/api/v1/knowledge/claims/687458MCP
get_claim(id=687458)