iFlytek SparkDesk (Spark 4.0 / Lite / Pro)
LLMby iFlytek
iFlytek's SparkDesk models are competitive in Chinese language tasks and offer the permanently free Spark Lite API. Known for strong speech and NLP capabilities built on iFlytek's long AI history.
Registration & API Key Steps
Step 2
Register with phone number and verify via SMS code.
Step 3
Complete real-name verification (may be required for full access).
Step 4
Create a new application — fill in the app name, category, and description.
Step 5
Wait for approval (usually within minutes).
Step 6
Obtain your three credentials: APP ID, API Key, and API Secret.
Pricing
| Tier | Price | Features |
|---|---|---|
| Spark Lite | Free | Permanently free API. Limited features. |
| Spark Pro | ¥3 / ¥3 per 10K tokens | Input / Output (~$0.41). Professional model. |
| Spark 3.5 Max | ¥2.1 / ¥2.1 per 10K tokens | Input / Output (~$0.29). Best value for advanced tasks. |
| Spark 4.0 Ultra | ¥5 / ¥5 per 10K tokens | Input / Output (~$0.69). Most capable model. |
Application Tips
Tip 1
Spark Lite is permanently free — one of the best free LLM APIs for Chinese tasks.
Tip 2
Spark 3.5 Max at ¥2.1/10K tokens is significantly cheaper than Baidu and Alibaba's top models.
Tip 3
Use company email for faster application approval.
Tip 4
iFlytek also offers strong speech recognition and synthesis APIs.
Tip 5
WebSocket API is the primary interface — different from typical REST APIs.
Tip 6
Authentication requires HMAC-SHA256 signature — use official SDK for easier integration.
China Access Solutions
Access Solution
Directly accessible in China. iFlytek is a major Chinese AI company. Supports Chinese phone, domestic payment methods.
Code Example
# SparkDesk uses WebSocket API with HMAC-SHA256 authentication
# Recommended: use the official Python SDK
# pip install sparkdesk-api
from sparkdesk_api.core import SparkAPI
sparkAPI = SparkAPI(
app_id="your-app-id",
api_secret="your-api-secret",
api_key="your-api-key",
version=3.5 # Spark version
)
response = sparkAPI.chat("请介绍一下量子计算")
print(response)
# --- Using WebSocket directly requires HMAC auth ---
# See official docs: https://www.xfyun.cn/doc/spark/Web.htmlRate Limits
| Tier | Limits |
|---|---|
| Default | Spark Lite (free): 2 concurrent connections. Spark Pro: varies by plan. Paid models support higher concurrency. |
Recommended Use Cases
Related API Guides
OpenAI GPT-4o / GPT-4.1 / o3
OpenAI
OpenAI's flagship LLM family including GPT-4o for multimodal tasks, GPT-4.1 for long-context coding, and o3 for advanced reasoning. Industry-leading models with the largest developer ecosystem.
Anthropic Claude (Sonnet 4.5 / Opus 4.5)
Anthropic
Anthropic's Claude model family excels in nuanced reasoning, safety, and long-context tasks. Claude Sonnet 4.5 offers the best balance of cost and performance, while Opus 4.5 delivers frontier intelligence.
Google Gemini (2.5 Pro / 2.5 Flash)
Google's Gemini models offer a generous free tier, 1M token context window, and strong multimodal capabilities. Gemini 2.5 Pro leads in reasoning, while Flash models provide cost-effective alternatives.
Meta Llama 4 (Scout / Maverick)
Meta
Meta's open-source Llama 4 models are free to use and available through multiple cloud providers. Llama 4 Scout and Maverick offer competitive performance at extremely low cost through partner APIs.