A January 16, 2026 publication presents Step Audio R1.1 as an interactive real-time spoken-dialogue model and describes its license and architecture.
On January 16, 2026, a publication presented Step Audio R1.1 as a model for interactive real-time spoken dialogue, under the Apache 2.0 license. According to the text, the model “thinks while it speaks” and uses a dual-brain architecture to reduce latency.
The publication highlights the model’s interactive speech scope and these claims about its license and architecture, but provides no evaluation results. To verify them, consult the original model page and check the license terms and technical documentation directly. If you use AI to study or apply the material, protect organizational data by following your organization’s privacy policy.