GPT-4o costs ¥60 per million input tokens on SSSAICode, while Claude 3.5 Sonnet runs ¥80 per million input tokens — both priced at roughly 1.2x the official API rates after the ¥100 minimum top-up.
Pricing Breakdown
SSSAICode doesn’t publish per-token rates publicly, but based on community data and the platform’s pricing tiers, here’s what you’re looking at:
| Model | Input (per 1M tokens) | Output (per 1M tokens) | Min Top-Up |
|---|---|---|---|
| GPT-4o | ¥60 | ¥200 | ¥100 |
| Claude 3.5 Sonnet | ¥80 | ¥240 | ¥100 |
| The ¥100 minimum recharge is the biggest blocker for casual testing. Compare this to direct OpenAI API access (which requires a foreign credit card and VPN) — if you’re already paying for a VPN, SSSAICode’s markup may actually be cheaper than the VPN + API combo. |
Cost Comparison vs Direct API
Direct OpenAI GPT-4o pricing (as of early 2026): $2.50/1M input tokens (¥18 at current rates). SSSAICode’s ¥60/1M is a 3.3x markup. Direct Claude 3.5 Sonnet: $3/1M input (¥21.50), so ¥80/1M is a 3.7x markup.
This is higher than most relay stations. The justification? You’re paying for the documentation and the Hong Kong/US node routing that gives you stable access without VPN.
Hidden Fees
No hidden fees that I’ve found. The price you see on the model selection page is what you pay. No batch discounts, no volume pricing tiers. The “premium pricing on MAX groups” mentioned in the cons refers to priority queues during peak hours — those cost extra but aren’t mandatory.
Pros & Cons
Pros:
- Documentation is genuinely excellent — best I’ve seen among Chinese relay stations. API reference, rate limit explanations, and error code guides are all in Chinese and English.
- Hong Kong and US nodes mean consistent latency (~200ms from mainland China to Hong Kong node).
- Active community on their Discord/WeChat group; developers actually get responses from the team. Cons:
- ¥100 minimum top-up is steep for someone who just wants to test with ¥10-20 worth of tokens.
- No free trial or demo credits. You’re committing ¥100 blind.
- Premium pricing on priority queues — if you need consistent low latency during Chinese peak hours (8-10 PM CST), expect to pay 20-30% more.
Verdict
SSSAICode is for developers who value stability and documentation over cost. If you’re building a production app that needs reliable GPT-4o or Claude 3.5 Sonnet access without VPN headaches, the ¥100 barrier is worth it — you’ll burn through that in a week of moderate use anyway. For casual users or those just experimenting: skip it. The markup is too high for toy projects, and the ¥100 minimum is a real commitment. Look at platforms with lower minimums or free tiers. The documentation alone saves hours of debugging. If that has value to you, SSSAICode delivers.
FAQ
Q: Can I use SSSAICode without a VPN from mainland China? A: Yes. That’s the entire point. The Hong Kong and US nodes handle routing so you don’t need a separate VPN. Q: Is there any way to try the service without paying ¥100? A: No. There’s no free trial, no demo credits, and no lower minimum recharge option. You pay ¥100 upfront. Q: Does SSSAICode support batch processing discounts? A: No. Pricing is per-token with no volume discounts. The rates stay the same whether you process 1 request or 1 million. Q: What happens if my account runs out of credits mid-request? A: The request fails and you need to top up. There’s no overdraft or credit buffer. Keep at least ¥20-30 in your account if you’re running production workloads. Q: How does SSSAICode compare to direct OpenAI API pricing in 2026? A: SSSAICode charges roughly 3.3x for GPT-4o and 3.7x for Claude 3.5 Sonnet compared to direct API rates. You’re paying the premium for no-VPN access and the documentation quality.
Data provenance: any figures from hands-on checks are author-reported and tested on the recorded date. Independent verification is unavailable unless a source is linked.
