Today, we’re proud to unveil a pioneering partnership with @Groq. Our Dialog voice model now runs on GroqCloud. This isn't just an upgrade; this is a leap forward that redefines what voice technology can be—fast, powerful, and beautifully human.

Google is taking local AI to the next level with Gemma 3 QAT GGUF models Uncompromised quality with quantization-aware training and uncompromised on-device performance with ggml. This is the way!

We are setting a new speed standard. At 219 characters per second, we're over 3x faster than the closest competitor. Speed isn't just a metric—it’s what powers real-time, scalable voice experiences across industries. Powered by Groq. Built for performance. The future of voice

I am having a lot of fun with this - e.g. record myself or girlfriend in Bulgarian and change to whatever.

bit.ly/4k2yCOj Hey, we at Play AI are searching for talented Inference engineers. Come work with us if you want to: - Build the conversational voice interface of the future - Push our models to achieve the highest quality and speed in the space - Work with cutting-edge

PlayAI Am I doing this right?

Diffusion in-painting and TTS model. Give it a try! We are expecting you to build some great things with it. Let me know if you need any help.

Another example of the multiple TTS parallel pipelines pattern. Here's a voice AI agent that speaks both English and Arabic, using a specific model/voice for each language. These are PlayAI voices. The STT, TTS, and LLM inference is all running on Groq Inc. (The LLM is

Just getting started

Llama.cpp supports the new gpt-oss model in native MXFP4 format The ggml inference engine (powering llama.cpp) can run the new gpt-oss model with all major backends, including CUDA, Vulkan, Metal and CPU at exceptional performance. This virtually brings the unprecedented quality

just remembered that time Noam Shazeer dropped the hardest line ever written in an ML paper