3 min read
Cartesia Sonic 3.6 pushes sub-90ms TTS across…
Cartesia's latest Sonic ranks #1 on Artificial Analysis voice leaderboards with 44-language coverage and agent-grade latency. What changed for production voice stacks.
5 min read
VoxCPM2: fine-tune a custom voice in 10…
OpenBMB's VoxCPM2 is a 2B tokenizer-free TTS model with Apache 2.0 weights, 30 languages, and LoRA fine-tuning from 5-10 minutes of audio. Here is how it compares to hosted APIs for voice agent projects.
5 min read
MisoTTS open-weights an 8B voice model claiming…
Miso Labs released an emotive 8B TTS with RVQ audio tokens and optional audio context. The 110ms number is hosted TTFB on big GPUs. Here’s what matters if you build voice agents.
4 min read
Fish Audio S2.1 Pro clones a voice…
Fish Audio raised $52M on $21M ARR with a voice model that clones from a 5-second clip and streams first audio in about 70ms. Here is what matters if you ship voice agents.