# StepFun Launches Real-time Voice Large Model StepAudio 2.5 Realtime

- Category: Flash (快讯)
- Published: 2026-05-09T12:00:00.000Z
- Source date: 2026-05-09
- HTML: https://www.88z.ai/en/news/flash-20260509-26dc5805c8fe22eb
- Markdown: https://www.88z.ai/en/news/flash-20260509-26dc5805c8fe22eb.md
- Other language: https://www.88z.ai/news/flash-20260509-26dc5805c8fe22eb.md

StepFun has released the new generation real-time voice large model StepAudio 2.5 Realtime, featuring three core advantages: top-tier paralinguistic capabilities, millions of customizable personas, and leading conversational EQ and IQ.

## At a glance

- Supports millions of customizable personas with fine-tuned personality and catchphrase adjustments
- Features top-tier paralinguistic capabilities for precise perception of tone and emotional details

## The update

StepFun has officially launched the new generation real-time voice large model, StepAudio 2.5 Realtime. This model achieves significant breakthroughs in interaction experience, primarily demonstrated through three core capabilities. First, it possesses top-tier paralinguistic processing abilities, accurately perceiving and reproducing subtle emotional details such as tone, speech speed, and pauses, making communication more engaging. Second, the model supports customization of millions of personas, allowing users to finely adjust traits from personality characteristics to catchphrases, thereby creating unique exclusive AI characters. Furthermore, StepAudio 2.5 Realtime has advanced comprehensively in conversational intelligence, achieving leadership in both emotional and intellectual quotient, ensuring interactions are both smart and considerate.

## Sources

- [阶跃星辰](https://mp.weixin.qq.com/s/P0x9EC_L_RRfSU6BQfyckA) (2026-05-09)

This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.