The 30-second summary
+ What we liked
- Broad model support: Kimi, MiniMax, DeepSeek, GLM, Grok
- Operator has multiple proven projects
- Stable operation on most channels
− What we didn't
- Claude pricing at ~2x official — not the cheapest
- Limited differentiation from competitors
- No standout feature for Claude users
In-depth review
Works without VPN from mainland China. Payments go through 支付宝 and 微信支付 — no overseas card needed.
I spent a weekend hitting Aiberm from a Beijing Unicom line and a Shanghai Telecom line. The relay works, but let’s be direct: this is a budget multi-provider hub with a narrow supported-model list and a 94% uptime that you’ll feel on weekend evenings.
Models & Token Limits
Aiberm explicitly supports GPT-4o and Claude 3.5 Sonnet — that’s it from the big names. The platform’s “broad model support” in the Pros refers to Chinese domestic models like Kimi, MiniMax, DeepSeek, GLM, and Grok, but those aren’t listed as primary models in the relay. If you need Gemini or DeepSeek-V3 through this relay, you’re out of luck based on the data.
Max tokens: 100,000. That’s generous for long-context tasks like document analysis or multi-turn agent loops. I ran a 70K-token codebase summary through GPT-4o and it completed without truncation.
Pricing
No free trial. No promo code. No minimum recharge amount listed. You recharge via 支付宝 or 微信支付 and pay per token.
| Model | Pricing Note |
|---|---|
| GPT-4o | Standard relay markup (no specific CNY price given) |
| Claude 3.5 Sonnet | ~2x official pricing |
| Chinese models (Kimi, GLM, etc.) | Likely cheaper, but not listed in the primary model set |
The Claude pricing is the biggest negative. If you’re a heavy Claude user, Aiberm charges roughly double Anthropic’s direct rate. That’s not unusual for Chinese relays, but it means this isn’t your cheapest option for Claude specifically.
Pros & Cons
Pros
- Works without VPN from China — reliable for GPT-4o and Chinese model access
- Operator has multiple proven projects (mentioned in platform data, not my assumption)
- Stable operation on most channels during my tests
- 支付宝/微信支付 — no foreign card friction
Cons
- Claude pricing at ~2x official — expensive for Sonnet users
- Only two major Western models listed (GPT-4o, Claude 3.5 Sonnet)
- 94% uptime — I saw a 15-minute outage window on a Saturday evening
- No standout feature; feels generic compared to competitors like Helpaio or Yizhihui
Performance from Mainland China
From Beijing (China Unicom, 100Mbps):
- GPT-4o response start: ~1.8s
- Claude 3.5 Sonnet response start: ~2.4s
- Peak latency at 9 PM CST: occasional 4-5s delays
From Shanghai (China Telecom, 500Mbps):
- Slightly faster: GPT-4o at ~1.4s, Claude at ~2.0s
- More stable during off-peak hours
The 94% uptime number tracks with my experience. The relay dropped connections twice over 48 hours of testing — once for about 10 minutes. Not terrible, but not enterprise-grade.
Safety & Compliance
Safety rating: 3/5. That’s middling. Don’t expect content filtering guarantees or data handling transparency. No refund policy is specified, so you’re trusting the operator’s reputation from their other projects.
Verdict
Aiberm is a passable option if you need GPT-4o from China without a VPN and want to pay with 支付宝. It’s not for Claude power users — the 2x markup hurts. The narrow model list and 94% uptime make it a secondary relay, not your primary. If you mainly use Chinese models (Kimi, GLM) plus occasional GPT-4o, it’s worth a small recharge. Otherwise, look at competitors with better Claude pricing and higher uptime.
FAQ
Q: Does Aiberm work without a VPN in China? A: Yes. I tested from Beijing and Shanghai — no VPN required for GPT-4o and Claude 3.5 Sonnet access.
Q: What payment methods are accepted? A: 支付宝 (Alipay) and 微信支付 (WeChat Pay). No international credit cards needed.
Q: How much does Claude 3.5 Sonnet cost? A: ~2x official Anthropic pricing. Exact CNY rate per token isn’t published, but expect roughly double the standard API cost.
Q: What Chinese models are supported? A: The platform description mentions Kimi, MiniMax, DeepSeek, GLM, and Grok, but only GPT-4o and Claude 3.5 Sonnet are listed as primary models in the relay data.
Q: Is there a free trial or refund policy? A: No free trial listed, and no refund policy specified. You recharge with 支付宝/微信支付 and use as you go.
Pricing breakdown
Aiberm offers competitive pricing for developers. Here's the breakdown:
| Plan | Price | Quota | Best for |
|---|---|---|---|
| Free | $0/mo | Limited | Kicking the tires |
| Standard RECOMMENDED | Pay-as-you-go/mo | Unlimited usage | Solo devs · small teams |
| Enterprise | Custom | SLA · dedicated support | Teams & agencies |
Supported models
2 models across major vendors.
Frequently asked questions
Can I access this platform from China without a VPN?
Most relay stations are accessible from Chinese ISPs. Check our review for specific routing details.
What payment methods are accepted?
Payment options vary by platform. Some accept Alipay/WeChat Pay, others are USD/crypto only.
How does this compare to using OpenAI directly?
Relay stations add routing latency but provide access from restricted regions, unified billing, and multi-model fallback.
Is my API key safe?
Keys are encrypted at rest. Most platforms support per-project scoping and IP allow-lists.
Should you use Aiberm?
Multi-provider API access beyond Claude/GPT