Mid-tier Mainstream models Direct connect ★ 3.8 / 5

LinkAi Review: Pricing & Comparison

Clear documentation, high bonus ratio on large top-ups — ¥100 gets you ¥30 free — covers both Claude and GPT

Last verified: 2026-08-11 · Visit official site →

Doing the math on the “bonus”

LinkAi’s (linkai.shop) core selling point is the bonus ratio on large top-ups — top up ¥100, get ¥30 free; top up ¥500, get ¥150 free. At first glance it looks like a marketing gimmick, but once you do the math, the discount is real.

Take Claude Sonnet 4.6 (official price $3/M input tokens) as an example:

Using the official API directly, no top-up bonus (at an exchange rate of ¥7.2/USD):

  • Input: ¥21.6/M tokens
  • Output: ¥108/M tokens

Top up ¥100, get ¥30 free (¥130 actual credit, but you only spend ¥100):

  • Effective discount: 100/130 ≈ 77%
  • Equivalent to paying about ¥5.5 per dollar of usage instead of ¥7.2 — a 23% saving

For users spending more than ¥300/month on the Claude API, a 23% discount is a meaningful saving. The catch is you need to be willing to prepay — the bonus only applies at the moment of top-up, you can’t use first and pay later.

Documentation: an underrated part of the experience

At a lot of relays, the documentation is written as an afterthought — often incomplete or out of date. The assessment that LinkAi’s documentation is relatively thorough comes from actual feedback shared by community users.

What does “thorough documentation” actually mean here?

  • Clear instructions for getting your base_url and API key
  • Configuration examples for mainstream tools like Claude Code and Cursor
  • Common error codes and how to resolve them
  • Model name mapping (relay model names sometimes don’t exactly match the official ones)

For a developer configuring a mainland Claude relay for the first time, documentation quality is often what determines whether the setup works within an hour or not.

Top-up strategy recommendations

Given LinkAi’s bonus mechanism, here are a few top-up strategy suggestions:

Testing phase (recommended):

  • Start by testing the integration with the ¥5 signup bonus
  • Confirm the main Claude/GPT models call correctly
  • Test actual latency and stability over 1-2 days

Small-to-medium usage (¥100-300/month):

  • Go with the ¥100-for-¥30-bonus tier, getting ¥130 in actual credit each time
  • Don’t top up too much at once — observe stability over a few months first

Medium-to-high usage (¥300+/month):

  • The ¥500-for-¥150-bonus tier is the better choice, with ¥650 in actual credit
  • Or top up in ¥100 batches instead — same effect, but lower and more controllable risk

Model coverage: staying focused

LinkAi mainly covers the Claude and GPT families, without trying to be a sprawling all-in-one aggregator:

Claude lineup:

  • Claude Fable 5 (latest flagship)
  • Claude Opus 4.8
  • Claude Sonnet 4.6
  • Claude Haiku 4.5

GPT lineup:

  • GPT-5.5
  • GPT-4o
  • The o-series reasoning models

This coverage already meets most individual developers’ day-to-day needs. If you also need DeepSeek, Gemini, or similar, you can pair it with other platforms — SiliconFlow covers open-source and domestic models, and FlintAPI focuses on 43 domestic large models.

Integration example

from anthropic import Anthropic

client = Anthropic(
    api_key="your LinkAi API Key",
    base_url="https://api.linkai.shop"  # check the official site for the current endpoint
)

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=2048,
    messages=[{"role": "user", "content": "Help me write a React component"}]
)

Claude Code configuration: set ANTHROPIC_BASE_URL to LinkAi’s endpoint and ANTHROPIC_API_KEY to your LinkAi key in settings, and you’re done.

Bottom line: LinkAi’s bonus mechanism offers real savings for medium-frequency users spending ¥100+/month, and its thorough documentation lowers the barrier for first-time setup. If you mainly use Claude + GPT and are comfortable with prepaying, LinkAi is worth considering.

Information verified 2026-08-11. The top-up bonus ratios could not be re-verified on the official site — check linkai.shop for current terms.

  • 01.AI (Lingyi Wanwu): founded by Kai-Fu Lee, dual-track open/closed-source Yi series, strong multilingual capability
  • Modal: on-demand serverless GPU, minimalist deployment for custom model inference
  • nexos.ai: founded by the Nord Security team, €30M Series A, enterprise security-compliant AI gateway
  • Eden AI: a single API aggregating OpenAI/Anthropic/Google, with built-in fallback routing

Quick facts

Pricing modelPay-as-you-go (third-party monitoring measures ~¥1–2/M tokens input); the top-up bonus ratios (¥100→¥30 etc.) could not be verified on the official site; covers Claude and GPT main models
Model coverageFull Claude lineup, full GPT lineup
Latency / SLAMainland direct connect, no proxy needed; check the official site for specific SLA
Mainland direct connectDirect connect
Best forDevelopers
Referral programNo public affiliate program found so far.

Pros

  • High bonus on large top-ups: top up ¥100 get ¥30 free (¥130 actual credit), top up ¥500 get ¥150 free (¥650 actual credit) — an effective discount of about 23%
  • Clear documentation: onboarding docs and usage instructions are relatively thorough, lowering the barrier for first-time setup
  • ¥5 signup bonus: test the integration at zero cost before deciding whether to top up

Cons

  • Model coverage is relatively narrow (mainly Claude + GPT), not suited to use cases that need multi-model aggregation
  • The bonus credit requires a top-up to unlock, not ideal for users who only want a small test
  • Long-term operating stability needs to be judged from community feedback — public information is limited

Compare more AI API relays

See the full comparison board — filter by price tier, model coverage, and mainland direct-connect status.

Back to the comparison board →