tools
iFlytek Spark (Xinghuo)
iFlytek Spark is an AI assistant for answering questions, reading documents, writing, coding, image generation, and multimodal chat.
iFlytek Spark provides conversational AI for search and question answering, writing, document and webpage summaries, translation, mathematics, coding, image generation, and voice or video interactions. Users can also browse or create task-specific agents.
It is available through web, desktop, mobile, browser-extension, and API interfaces. Consumer access is available free of charge, while API and model services can use token-based pricing. It does not replace professional medical, legal, or financial advice.
Features
- Summarizes webpages and uploaded Word, PDF, and TXT documents
- Generates workplace copy, lesson plans, papers, scripts, and presentations
- Solves equations, geometry problems, and probability questions
- Answers questions using web search and referenced sources
- Generates images from text prompts
- Generates, explains, and corrects code
- Supports image, audio, video, voice, and video-call interactions
- Creates and uses task-specific AI agents
Use cases
- Summarize research papers, reports, webpages, and meeting materials
- Draft resumes, emails, articles, scripts, lesson plans, and presentations
- Solve and explain mathematics homework and technical questions
- Translate foreign-language webpages, documents, images, or video frames
- Generate images for concepts, avatars, designs, and social content
- Build applications through the Spark API or customized agents
Pros
Cons
Pricing
- Starting price
- 输入：1元/百万tokens
- Pricing checked
- 2026-09-19
Spark-X2.5
输入：3.2 1.6元/百万tokens；输出：12 6元/百万tokens
- 256K context
- MoE architecture
- Supports more than 200 languages
- Code and agent capabilities
Spark-X2.5-1.7B
0元/百万tokens
- 1M context
- 1.7B lightweight edge model
- Supports more than 200 languages
- Limited-time free
Spark-X2.5-4B
0元/百万tokens
- 1M context
- 4B edge model
- Supports more than 200 languages
- Limited-time free
Spark-X2-Flash
输入：1元/百万tokens；输出：2元/百万tokens
- 256K context
- Optimized for agent scenarios
- Code and agent capabilities
Spark X2
输入：3元/百万tokens；输出：3元/百万tokens
- 192K context
- General-purpose model
- Supports more than 130 languages