Speech AI

Adalat AI has opened Indic ASR models for three languages, giving builders new speech-recognition assets while leaving performance details to verify.

Wispr raised $280 million at a $2 billion valuation, funding new meeting tools and speech models as competition intensifies in AI dictation.

NVIDIA's reported NemotronLabs VoiceChat 11B targets open, real-time voice agents, but limited source evidence leaves benchmarks and deployment details unconfirmed.

OpenAI has outlined GPT-Live, a low-latency voice system designed for continuous conversation, offering builders a new approach to realtime AI interaction.

Smallest.ai has raised $13M to build low-latency voice AI for enterprise calls, betting small specialized models can make agents sound human.

OpenAI has introduced Presence, a managed platform for enterprise voice and chat agents aimed at reliable, policy-controlled deployments.

Current AI is expanding a public-interest AI stack with offline language tools, grants, and an open-source chatbot aimed at underserved communities.

OpenAI has introduced GPT-Live voice models for ChatGPT, betting that full-duplex speech can make AI assistants more useful for live work.

Interfaze has released diffusion-gemma-asr-small, an open-source ASR model for six languages that tests diffusion decoding as a new path for speech AI.

Google released Gemini 3.5 Live Translate, a speech-to-speech AI model for near real-time translation in over 70 languages.