iFlytek releases the iFlytek Spark X2.6 cloud-device integrated large model

A large model trained on fully domestic chips is here! Million-token context, priced lower than DeepSeek V4.1 Flash

Zhidongxi Download the app for exclusive early access October 22, 2026, 21:35:27 From Beijing

Zhidongxi reported on October 22 that today iFlytek released several new large models and AI application achievements, including the device-cloud integrated large model iFlytek Spark X2.6, the speech foundation model SparkAudio 1.0, the Spark embodied multimodal interaction technology, the WinWork digital employee platform, and the super AI teaching assistant iFlytek Starlight.

Its device-cloud integrated iFlytek Spark X2.6 foundation model completed a capability upgrade based on fully domestic computing power, including the cloud-side 293-A30B model and the device-side X2.6-4B model, the latter of which has been open-sourced.

The cloud model 293-A30B supports a million-token context, with an input price of 0.9 yuan per million tokens, an input cache hit price of 0.08 yuan per million tokens, and an output price of 3.6 yuan per million tokens. Its input cache miss and output prices are lower than those of DeepSeek-V4.1-Flash-1017.

According to test results published by iFlytek, the Spark X2.6-4B model surpasses its previous generation Spark-X2.5-4B as well as the Qwen3.5-4B, Gemma4-E4B, and Qwen3.5-9B models in information search Q&A, office document processing, schedule and task management, local spreadsheet operations, cross-tool collaboration, multi-turn tool calling, and complex knowledge search.

In addition, iFlytek also officially announced several AI hardware products, such as the iFlytek AI recording ring, iFlytek AI glasses, and the iFlytek AI PC series. Among them, the Lingji N9000 AIPC laptop has up to 160 TOPS of local AI computing power; the Spark Office All-in-One machine provides 2240 TOPS of AI computing power and supports private deployment of large models.

At present, iFlytek's product matrix has the iFlytek Spark foundation model as its base and has achieved applications in fields such as education, healthcare, education, law, automobiles, and robotics.

He believes that a fully self-controllable full-stack large model is the inevitable choice for AI security and extreme cost-effectiveness, and this requires a fully self-controllable full-stack device-cloud integrated large model to solve. In the future, 60% of tasks will run locally, 30% of tasks will be processed through cloud-device collaboration, and 10% of tasks will be processed independently in the cloud. This is also the key reason iFlytek launched the device-cloud integrated large model X2.6 this time.

1. Trained on fully domestic computing power, debuting a device-cloud integrated large model

In the benchmark test results disclosed by iFlytek, in terms of coding capability, Spark X2.6 took first place in the SWE Pro and SWE Multilingual test sets, and ranked second only to DeepSeek-V4.1-Flash 1017 in the SWE Verified test set; in the agent track, it is in the first tier in both enterprise operations and maintenance Agents and the general tool-calling MCP-Atlas; in terms of reasoning capability, Spark X2.6 is on par with MiniMax M3 in the AA-LCR v1.1 test set.

On the training side, the model broke through the long-horizon task training bottleneck of domestic chips. Taking the Feixing-2 domestic ultra-large-scale intelligent computing platform's ten-thousand-card training on 2000B 256K-length data as an example, the training time was shortened from 4.5 months to 1.2 months, and ultra-long-context training efficiency increased to 4 times the original. Liu Qingfeng said that Spark X2.6 was trained in the past 2-3 months.

At the inference level, based on distributed inference technology for domestic heterogeneous computing power and the first four-layer KV Cache architecture on fully domestic hardware, iFlytek overcame problems such as insufficient capacity, high inference latency, and low cache hit rates in long-context scenarios. The inference efficiency of Feixing-2 domestic chips reached 4.5 times that before optimization, and the KV Cache hit rate increased from 63% to 95.7%.

On the device-side model front, iFlytek created the Spark X2.6-4B model. Liu Qingfeng said this is because the 4B size is very suitable for deployment on robots, automotive platforms, and AI PCs. Since its open-sourcing on September 1, the previous generation Spark X2.5-4B device-side model has reached 2.56 million downloads, ranking among the top four device-side models globally.

To improve device-side model capabilities, iFlytek adopted device-cloud integrated training technology. Spark uses an adaptive iterative post-training algorithm that perceives device-side model capabilities, dynamically selecting teacher models and adjusting training weights according to device-side capabilities.

Based on a "cloud + device" collaborative architecture with precise routing by the device-side model, the device-side model first determines how a task should be completed: purely on-device, purely in the cloud, or through device-cloud collaboration. A key issue here is whether device-cloud routing is accurate. iFlytek introduced a device-cloud routing collaborative training mechanism based on a critique model, raising device-cloud routing accuracy from 81.2% to 96.1%.

On top of the device-cloud integrated Spark X2.6 model, iFlytek launched two products: Spark Loomy for daily office work, helping users write reports, make PPTs, and analyze data so daily work can be completed easily; and Spark AStudio for professional office work, helping users write code and conduct research so complex work can advance autonomously.

During the live demo, a presenter played the role of a startup's general manager, showing how Spark AStudio can, without an internet connection, help the user review recent weekly reports and emails from major regions, then analyze and organize the business situation. After retrieving the content, the model generates visualized tables, three to-do items, risks, and more, and also provides corresponding email links.

In tasks that involve collaboration with the cloud model, the model first sorts through the local data, and only when a cloud collaboration task is needed does it initiate a collaboration request. Before that, the local model desensitizes the data. After the cloud finishes processing, the local model processes the previously desensitized information again to generate complete content.

In terms of bringing AI from the digital world into the physical world, iFlytek released the SparkAudio 1.0 speech foundation model and the new AI hardware iFlytek AI Recording Ring, among others.

The SparkAudio 1.0 speech foundation model is trained on fully domestic computing power. At the event, SparkAudio 1.0 analyzed background music, male voices, mechanical operating sounds, bicycle bells, and many other sounds in a complex audio clip.

Behind this model are three technological innovations: a semi-supervised audio encoder that brings a 20% performance improvement, dynamic segmented end-to-end modeling that reduces response time by 15%, and dual-paradigm collaborative decoding that reduces cost by 60%.

The model has already been launched in products such as iFlytek Input 16.0, iFlytek Hearing, the iFlytek Smart Office Book, and iFlytek Translator. Among them, after the upgrade, the iFlytek Smart Office Book supports handwritten annotation of speaker names, with subsequent speech automatically matched to them; it can also distinguish speakers when people interrupt or multiple people speak over one another; and it recognizes mixed Chinese-English speech and technical terminology.

iFlytek also released a new AI hardware product, the iFlytek AI Recording Ring, weighing 35g and designed to magnetically attach to a phone. Users only need to tap the recording button on the ring to record anytime, anywhere. After recording ends, it transmits at high speed via WiFi 6, sending about 10 seconds of a 2-hour recording to the phone and converting it to text in about 2 minutes; it can also intelligently distinguish speakers, generate AI meeting minutes, extract key points and to-do items, and support multilingual simultaneous interpretation.

The device is now on sale on e-commerce platforms, priced at 1,299 yuan.

In multimodal interaction, iFlytek has mainly improved the ability of smart terminals such as cockpits and glasses to understand needs.

The iFlytek AI Glasses have display capability, weigh 40g, support 90dB exhibition noise suppression, and pick up sound from up to 8 meters away. The device can translate more than 120 languages, connect to PCs and cloud agents, and support remote command of office work on a computer. The device starts at 4,299 yuan.

The iFlytek Spark Smart Cockpit SparkAuto Space 3.0 has greatly improved in multi-person interaction response rate, multi-topic interaction accuracy, outside-vehicle far-field interaction distance, and success rate of complex agent tasks.

Embodied multimodal interaction for robots faces challenges such as noise-reduction recognition for robot movement in open spaces, audio-visual spatiotemporal alignment of sound sources and multiple interacting people, and coordination of multimodal interaction and action linkage.

Based on this, iFlytek released the Robot Super Brain Platform V3.0, which can achieve dynamic noise reduction in complex spaces, interaction intent recognition for multiple target people, unified planning and execution of speech and actions, and vision-triggered proactive interaction.

In a "7×24-hour" AI unmanned supermarket scenario, the robot can combine speech, vision, and location information to identify the needs of different customers, and even when multiple people interrupt or user needs change repeatedly, it can complete tasks such as picking up goods, delivering them, and scanning codes for checkout.

The Robot Super Brain Platform V3.0 will be opened to developers, including the Spark Super Brain Board, multimodal interaction modules, embodied intelligence development kits, and the embodied multimodal interaction Harness.

Based on this, iFlytek launched the iFlytek Xingfan WinWork platform. The platform can, by building an enterprise business ontology, accumulate an enterprise's business procedures, business data, and expert knowledge base, so that digital employees better understand the business; WinWork can also be embedded in enterprise business processes to support digital employees in participating in a complete work loop, from autonomously identifying tasks, breaking them down and planning, intelligently distributing them, to linking various business systems for execution.

In addition, iFlytek launched the iFlytek AI PC series products based on the Spark device-cloud large model and domestic operating systems. Among them, the Lingji N9000 AIPC laptop has up to 160 TOPS of local AI computing power; the Spark Office All-in-One provides 2,240 TOPS of AI computing power and supports private deployment of large models.

iFlytek Xingguang integrates lesson preparation, teaching research, home-school communication, and administrative assistance into a single workbench, supporting the creation of lesson plans, animations, and scenario videos, multi-role simulated roundtable discussions, standardized form filling, statistical analysis, and visual presentation. Internally, it connects to channels such as Zhixue.com and Xiaoxin; externally, it connects to remote control channels on multiple platforms such as WeCom, QQ, Feishu, and DingTalk.

▲Helping teachers fill out home-school communication forms

In a lesson preparation scenario, to help students understand the relationship between current direction and magnetic pole direction, a teacher submits a request to iFlytek Xingguang, and the system generates a 3D interactive resource that can be rotated for observation; it can also be combined with camera-based gesture operation so that students can simulate the right-hand grip rule.

▲A 3D interactive resource that can be rotated for observation

In the medical field, iFlytek Imaging Cloud is a global medical imaging AI service platform covering regions with more than 260 million people nationwide. The medical imaging cloud platform supports image sharing, intelligent assisted diagnosis, quality control, and report generation; iFlytek Xiaoyi is an AI health assistant by residents' side, having completed more than 220 million AI consultations cumulatively, with downloads exceeding 44 million.

The X2.6 device-cloud integrated large model unveiled this time is centered on finely splitting and scheduling AI task requirements, thereby balancing the contradictions among performance, cost, and data security, and offering developers and enterprises a new option for deployment. Especially for industries that are highly sensitive to data leaving their premises, such as education, healthcare, and law, this architecture can meet their needs for privatization and local deployment.

In addition, competition in today's large model industry is continuing to extend toward computing power adaptation, scenario deployment, data security, and integrated software-hardware delivery. In the future, pure foundation model players that only have model capabilities but lack industry closed loops and hardware carriers may be eliminated.

"Special statement: The above content (including videos, pictures, and audio if any) is uploaded and published by users of "Dafeng Hao," a self-media platform under Phoenix Net. This platform only provides information storage space services.

Notice: The content above (including the videos, pictures and audios if any) is uploaded and posted by the user of Dafeng Hao, which is a social media platform and merely provides information storage space services."