iFlytek releases SparkAudio 1.0 speech large model and Robot Super Brain Platform V3.0

Related topics top start Related topics top end Futures promotion begin Futures promotion end Miaopai begin Miaopai end Video player start Video player end Market chart begin Market chart end

Sina Tech news October 23 evening: At the main forum of the 2026 iFLYTEK Global 1024 Developer Festival, Chairman Liu Qingfeng released the SparkAudio 1.0 speech large model, multimodal interaction technology, and Spark embodied multimodal interaction technology and related products, announcing that it will continue to advance along "voice interaction—multimodal interaction—embodied multimodal interaction," leading the implementation of a new generation of human-computer interaction with integrated software-hardware technology and product innovation.

In the field of voice interaction, high-noise recognition has long been an industry difficulty, and complex audio perception and multi-person speech analysis on that basis are even more advanced functions. Addressing these difficulties, iFLYTEK released the SparkAudio 1.0 speech foundation large model, bringing comprehensive upgrades to intelligent voice capabilities including speech recognition, simultaneous interpretation and interaction, as well as audio understanding, emotion recognition, real-time speaker separation, and environmental sound understanding.

It is reported that SparkAudio 1.0 has capabilities similar to a "sound detective" in crime dramas: even on a noisy street, it can distinguish horn sounds, street vendors' calls, and the sound of wheels rubbing against the ground, separate overlapping human voices, and infer the surrounding environment from sound clues—not only "hearing clearly what is said" but also "understanding what is happening." Related speech recognition tests show an overall effect improvement of 20%, a response time reduction of 15%, and a cost reduction of 60%, making voice services more accurate, more timely, and easier to scale.

Currently, SparkAudio 1.0 has launched in iFLYTEK products; users can experience it in products such as iFLYTEK Input Method 16.0, iFLYTEK Hearing, and iFLYTEK Smart Office Book. At the event, the iFLYTEK AI Recording Ring was also released; it can magnetically attach to the back of a phone like a phone stand, start recording at any time, transcribe 2 hours of recording in only 2 minutes, and has a comprehensive recognition accuracy rate exceeding 96%.

In addition, iFLYTEK also released multimodal interaction technology, further integrating multimodal far-field recognition, multi-person interaction, audiovisual fusion understanding, personalized memory, and emotional semantics, enabling AI to understand richer scene information. iFLYTEK AI Glasses are a typical application product of this technology. At the event, Lin Huijie, General Manager of iFLYTEK's Wearable Business Division, demonstrated on site the summary of Liu Qingfeng's speech process just recorded with iFLYTEK AI Glasses, with clear structure and rich text and images.

From intelligent terminals to mobile robots, interaction must also adapt to spatial changes, multi-person participation, and action execution. Addressing these challenges, iFLYTEK released Spark embodied multimodal interaction technology, achieving breakthroughs in dynamic noise reduction in complex spaces, interaction intention recognition for multiple target people, unified planning and execution of speech and actions, and vision-triggered proactive interaction, enabling robots to both communicate with people and collaboratively complete tasks in dynamic environments, providing key technical support for their entry into home and social service scenarios.

iFLYTEK also launched the Robot Super Brain Platform V3.0, integrating the Spark Super Brain Board, embodied intelligence development kit, multimodal interaction module, and embodied multimodal interaction Harness; through device-cloud collaboration and front-back office collaboration, it enables robots to chat continuously on the front end and keep handling tasks on the back end, and supports modular and configurable industry customization. (Wen Meng) news_keyword_pub,stock,sz002230&sz300024 iframe below main text iframe below main text General statement General statement QR code at end start QR code at end start

QR code at end start QR code at end start Editor name and work code

Editor name and work code