Yiye Zhiqiu API Review: Pricing & Comparison

A minimum top-up is a hidden barrier

Developers who’ve used mainland AI relays have probably run into this: you just want to test how well a given model performs, to see whether it fits your project — but the platform requires you to top up ¥20, ¥50, or more before you can even make a call.

For a small proof-of-concept test, that’s a disproportionate amount of friction. You don’t yet know how stable the relay is or whether the latency meets your needs, but you have to pay an “admission fee” first.

Yiye Zhiqiu API (88996.cloud) is positioned specifically to remove that barrier: no minimum top-up limit, top up only as much as you need.

Where Yiye Zhiqiu API fits best

Students and beginners: just starting to learn how to integrate an AI API, only need to get a few examples running, and don’t need to top up ¥100 for that.

Proof of concept (POC): evaluating whether some AI feature is viable, running a batch of test data on a few yuan worth of tokens, and confirming feasibility before considering formal integration.

Short-term, one-off projects: a one-time data processing task, or a temporary AI feature for the duration of an event — use it and be done, with no need to keep maintaining a balance on any given platform.

Multi-platform comparison: want to run an A/B comparison across multiple relays, topping up a few yuan at each one to run the same batch of tests and collect data cheaply.

How to connect

from openai import OpenAI

client = OpenAI(
    api_key="your Yiye Zhiqiu API key",
    base_url="https://api.88996.cloud/v1"  # check the official site for the current endpoint
)

response = client.chat.completions.create(
    model="gpt-5-5",  # check the official site for supported models
    messages=[
        {"role": "user", "content": "Explain in simple terms what an API is"}
    ]
)
print(response.choices[0].message.content)

The limits of a lightweight platform

Yiye Zhiqiu API’s lightweight positioning also comes with some limitations:

• Not suited to high-concurrency production: lightweight platforms typically lack dedicated rate-limiting guarantees and an SLA
• Less stable than established platforms: relatively limited operating history and user base
• Fairly basic feature set: don’t expect complex usage statistics, multi-key management, team collaboration, and similar features

Once your project moves from the validation stage into production, you should migrate to a platform with better stability and support, such as AiHubMix (a permanent free testing tier) or KoalaAPI.

Information verified 2026-07-04. Actual supported models and top-up methods are subject to the Yiye Zhiqiu API official site.

Related reviews

• ShiyunApi ⚠️ shut down: ⚠️ ceased operations in June 2026; the official site now redirects to TokenRiver (tokenriver.cn)
• UniAPI: OpenAI/Claude/Gemini plus multimodal options like Midjourney/Suno/Kling, mainland direct connect with no proxy needed, billed at per-channel discounts
• 35.AIGCBEST: focused on OpenAI relay, Azure-style pricing, exchange rate around 1.5¥/USD, stable and reliable
• GPTGOD: a reverse-engineered, ultra-low-price relay, exchange rate around 0.6¥/USD, extremely cheap but stability isn’t guaranteed