#StepFun AI

Step-Audio-EditX is an open-source 3B audio LLM that edits speech at the token level, enabling iterative, controllable changes to emotion, style, and paralinguistic cues through large-margin synthetic data and token-level PPO.

'StepFun AI released Step-Audio 2 Mini, an open-source 8B speech-to-speech model that combines unified audio-text tokenization, emotion-aware generation, and retrieval-augmented grounding to beat GPT-4o-Audio on multiple benchmarks.'