AI tool comparison
Gemini 3.1 Flash TTS vs VoxCPM2
Which one should you ship with? Here is the side-by-side panel verdict, pricing read, reviewer split, and community vote comparison.
Voice & Audio
Gemini 3.1 Flash TTS
Google's new TTS API: 70 languages, 200+ audio tags, native multi-speaker
75%
Panel ship
—
Community
Free
Entry
Gemini 3.1 Flash TTS is Google's new text-to-speech model, launched today on Google AI Studio and Vertex AI. It supports 70+ languages and introduces a natural-language audio tag system with 200+ expressivity controls — developers can describe delivery in plain English ("whisper conspiratorially", "warm and unhurried") and the model interprets those instructions at inference time. The model also supports native multi-speaker dialogue generation from a single prompt, outputting a conversation with distinct, consistent voices without requiring separate passes. All audio output is watermarked via Google's SynthID technology for provenance tracking. For developers building voice agents, podcasting tools, or multilingual apps, this is a meaningful upgrade over existing options. The audio tags approach in particular is a genuinely novel paradigm compared to prosody markup languages like SSML, and developer reception on X and HN has been strong — Simon Willison called out the expressivity controls as the standout feature.
Audio & Voice
VoxCPM2
Tokenizer-free TTS: voice design, cloning, and 30 languages from 2B params
75%
Panel ship
—
Community
Paid
Entry
VoxCPM2 is an open-source text-to-speech system from OpenBMB that takes a fundamentally different architectural approach to speech synthesis. Instead of the discrete tokenization pipeline used by most modern TTS systems, VoxCPM2 operates entirely in latent space through a diffusion autoregressive pipeline — bypassing tokenization altogether. The 2B-parameter model was trained on over 2 million hours of multilingual speech and supports 30 languages plus 9 Chinese dialects with no language tagging needed. What makes VoxCPM2 stand out is its three-mode voice control system. "Voice Design" lets you create entirely new voices from natural language descriptions alone — "young woman, gentle voice, slightly husky" — no reference audio required. "Controllable Voice Cloning" takes a reference clip and lets you adjust style and emotion. "Ultimate Cloning" provides maximum fidelity by supplying both the reference audio and its transcript. Output quality is 48kHz studio-grade audio, and the model runs at RTF ~0.3 on an RTX 4090 (or ~0.13 with Nano-vLLM acceleration). The Apache 2.0 license makes VoxCPM2 commercially viable for builders who've been held back by restrictive TTS licensing. It benchmarks competitively with commercial models on Seed-TTS-eval across English and Mandarin. The Hugging Face demo is live, weights are published, and it installs via `pip install voxcpm`. For any developer building voice products, this is worth evaluating immediately.
Reviewer scorecard
“This replaces ElevenLabs for a lot of use cases — and at Google's pricing it's hard to argue against. The natural-language audio tags are the real unlock: instead of wrestling with SSML prosody markup, you just describe what you want. The multi-speaker output from a single prompt is going to save a ton of orchestration code in voice agent pipelines.”
“Apache 2.0 + pip install + 48kHz output is the holy grail for voice product builders. Most open TTS models either sound robotic, have restrictive licenses, or require complex setup. VoxCPM2 clears all three bars. The voice design feature alone changes how you prototype voice UX — describe the persona instead of recording it.”
“It's Google — which means it could be deprecated in 18 months and replaced with Gemini 4 Flash TTS Pro Ultra. The audio tags sound creative but until there's a published spec for all 200+ of them, you're guessing at prompt-engineering your voice model. And SynthID watermarking is only as useful as the detection ecosystem, which is still nascent.”
“RTF of 0.3 on an RTX 4090 means real-time generation requires serious hardware — most small builders can't run this locally at scale. The technical report isn't published yet, so the benchmark claims are harder to independently verify. And 30 languages sounds impressive until you check whether your target dialect is actually well-represented in those 2M training hours.”
“Natural-language expressivity control for TTS is a paradigm shift. When the model can interpret 'sound like you're delivering devastating news gently' without explicit prosody markup, we're entering an era where voice synthesis becomes genuinely directorial. The 70-language coverage plus SynthID watermarking points toward a future where synthesized voice is both globally expressive and auditably provenance-tracked.”
“The shift away from discrete tokenization in TTS is architecturally significant — it mirrors the same trajectory that diffusion models took in image generation, and look how that ended. VoxCPM2 is an early signal that the tokenize-everything paradigm in audio is starting to crack. The end state is real-time, hyper-expressive voice synthesis running on consumer hardware.”
“I've been paying for ElevenLabs and manually tweaking prosody to get the right delivery. The audio tag system here could cut that iteration time dramatically — describing the scene and letting the model interpret is so much more intuitive than sliders and SSML. Multi-speaker from a single prompt is going to be huge for podcast generators and explainer video tools.”
“Designing voices with natural language instead of recording sessions is a genuine workflow unlock for content creators and game developers. The ability to describe 'tired, slightly gruff narrator in his 50s' and get consistent output is something I've wanted for years. The 48kHz output quality means it's usable in professional audio contexts without upsampling.”
Weekly AI Tool Verdicts
Get the next comparison in your inbox
New AI tools ship daily. We compare them before you waste an afternoon.