PANews reported that NVIDIA said RTX Spark now supports Nemotron 3.5 Transcribe and Liquid AI LFM3 series models, and has added AI resource-allocation functionality between foreground applications and background GPU agents.
Block has released the Dessa voice model, saying it was trained only on recordings of human conversations and supports full-duplex voice interaction. It can run locally on a MacBook Pro with an M4 Max or higher-spec configuration. The model's parameters, benchmark results, and weights have not yet been disclosed.
Block’s AI research lab released the research preview of the full-duplex voice dialogue model Dessa, saying it was trained solely on human conversational speech and can run locally on a high-specification MacBook. The model is not yet broadly available.
Beating AI reported that ElevenLabs released Eleven v4 and Eleven v4 Turbo. Citing the Artificial Analysis blind test rankings, the report said that Eleven v4 had 1315 Elo and was temporarily ranked first.