Publish time
-Model Series
Input type
Output type
Context Window
65,536Max Output Length
8,096Input Price
¥2 / 1M tokensOutput Price
¥3 / 1M tokensDeepSeek-V3.2-Exp 模型，是一个实验性（Experimental）的版本。作为迈向新一代架构的中间步骤，V3.2-Exp 在 V3.1-Terminus 的基础上引入了 DeepSeek Sparse Attention（一种稀疏注意力机制），针对长文本的训练和推理效率进行了探索性的优化和验证。
Zhinao API routes requests to the best-fit provider and automatically fails over to the one with highest availability.
TTFT
4.30s
Throughput
23.48tps
Uptime
100.00%
Provider Model
qiniu/deepseek-v3.2-exp
Supported Parameters
Recent Uptime
Reasoning
-
Supported Response Formats
Request Log Collection
-
Distillable
-
Total Context
65,536
Max Output
8,096
Input Price
¥2 / 1M tokens
Output Price
¥3 / 1M tokens
Compare different providers across Zhinao API
23.57 tok/s
4.34 s
Uptime for deepseek-v3.2-exp across all providers
Zhinao API normalizes requests and responses across providers for you
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.360.cn/v1",
apiKey: process.env.ZHINAO_API_KEY,
});
const response = await client.chat.completions.create({
model: "deepseek-v3.2-exp",
messages: [
{ role: "user", content: "Hello, how are you?" }
],
temperature: 0.7,
max_tokens: 1000,
});
console.log(response.choices[0].message.content);