Grok Voice Think Fast 1.0 vs VoxCPM2

xAI has launched Grok Voice Think Fast 1.0, its most capable voice model, now available via API. Positioned squarely at enterprise use cases — customer support, sales, and complex multi-step workflows — the model performs background reasoning without adding latency, letting it handle challenging queries while sounding like a natural conversation. At $0.05 per minute, it's priced aggressively against the market. The model's standout feature is structured data collection: it can accurately capture email addresses, phone numbers, street addresses, and account numbers even when spoken quickly, with strong accents, or with disfluencies. It supports over 25 languages and handles real-world messiness including noise, interruptions, and code-switching. This isn't a demo model — Grok Voice is already live powering Starlink's phone sales line (+1 888 GO STARLINK), where it converts 1 in 5 incoming sales inquiries into purchases. The launch puts xAI squarely in competition with ElevenLabs, Deepgram, and OpenAI's Realtime API. The Starlink deployment is a significant proof point that moves this beyond hype into production-grade enterprise voice AI.

VoxCPM2 is a 2B-parameter text-to-speech system from OpenBMB — the team behind MiniCPM — built around a tokenizer-free, diffusion-autoregressive architecture. Most TTS systems convert text to discrete audio tokens first, then decode those tokens to waveform. VoxCPM2 skips the tokenization step entirely, operating in continuous latent space. The result is 48kHz output with smoother prosody and finer pitch control than token-based systems. The headline feature is "Voice Design": you describe a voice in natural language — "a confident male voice, mid-Atlantic accent, slightly gravelly, deliberate pacing" — and VoxCPM2 synthesizes a brand-new voice from that description without any reference audio sample. This is architecturally different from voice cloning (which requires samples) and voice selection (which picks from a catalog). It supports 30 languages with automatic detection, no language tags required. The model runs on consumer hardware (~8GB VRAM), integrates with the MiniCPM-4 language model backbone, and is released under Apache 2.0. For developers building multilingual voice products or researchers exploring generative voice control, VoxCPM2 represents a meaningful step beyond current open TTS leaders like F5-TTS and CosyVoice.