DawCode Pricing: Input and Output Token Rates

GPT-4o input costs $0 per million tokens — but only for your first ¥4 worth of usage. DawCode is a Chinese relay station that gives developers ¥4 free credit (no credit card required) to test GPT-4o and Claude 3.5 Sonnet. It accepts Alipay and WeChat Pay, making it frictionless for local developers who want to bypass VPN costs. But the free tier has limits, and the paid pricing structure has quirks worth examining before you commit real money.

Pricing Breakdown: Free Trial vs Paid Plans

The free trial gives you ¥4 credit immediately upon registration. That’s enough to run roughly 40-80 GPT-4o requests depending on prompt length — decent for smoke-testing integration. No credit card required. No minimum top-up. Beyond the free credit, DawCode offers multiple route options per model. The “MAX” and “cc-stu” groups are both available. MAX packages generally provide higher daily quotas, while cc-stu packages have lower daily quotas relative to their cost. Recent price increases on cc-stu routes make them less attractive for heavy usage. Key takeaways:

• ¥4 free credit is genuinely no-strings-attached — no card needed
• cc-stu prices rose recently, reducing value for existing users
• No batch discount thresholds or volume pricing visible in the data

Cost Comparison: DawCode vs Direct API

Direct API pricing from OpenAI (2026 context): GPT-4o runs approximately $2.50 per million input tokens and $10 per million output tokens. Claude 3.5 Sonnet from Anthropic costs roughly $3 per million input tokens and $15 per million output tokens. DawCode’s relay pricing is not transparently listed per-token. Instead, it uses package-based quotas. This means you cannot directly compare per-million token rates without knowing the exact package breakdown. The lack of per-token pricing is a hidden overhead — you pay for a package, not for what you actually use.

• No per-token rates means you cannot calculate exact cost per request
• Package-based pricing may leave unused credits if you don’t hit quota limits
• Direct API is cheaper per-token if you can pay with a foreign card

Performance and Reliability

DawCode reports 100.0% uptime — which is suspiciously perfect for any API relay. Average latency sits at 1975ms (about 2 seconds per request), earning a speed rating of 3.0/5. That’s below average for relay stations, which typically aim for under 1 second. Safety rating is 4/5, meaning basic security practices are in place but no advanced encryption or audit logs are mentioned. The max token limit is 100,000 — sufficient for most chat applications but not for long-document processing. Key takeaways:

• 100% uptime claim is unverifiable and unusually high
• 2-second latency is slow; fine for batch jobs, bad for real-time chat
• 100K token limit covers most use cases but not heavy document workloads

Hidden Fees and Package Traps

Two specific cons from the data: “Some packages have low daily quota vs cost” and “cc-stu prices rose recently.” This means if you buy a cheap package, you might hit daily limits quickly and be forced to buy another. The recent price increase on cc-stu routes means returning users may pay more than new users. No refund policy is specified. No minimum recharge amount is documented. These gaps mean you should test with the free ¥4 credit before depositing any real money. Key takeaways:

• Package quotas may not match your actual usage patterns
• Price increases affect existing users — no grandfathering mentioned
• No refund policy means deposits are effectively non-refundable

Pros & Cons

Pros:

• ¥4 free credit with no credit card required
• Accepts Alipay and WeChat Pay — no VPN needed
• Multiple route options per model (MAX and cc-stu)
• Daily check-in rewards for active users Cons:
• Some packages have low daily quota relative to cost
• cc-stu prices recently increased
• Response speed below average (1975ms latency)
• No per-token pricing published — cost comparison is opaque
• No refund policy specified

Verdict

DawCode works as a low-commitment entry point for Chinese developers who want to test GPT-4o and Claude 3.5 Sonnet without a VPN. The ¥4 free credit is genuinely useful for integration testing. But the package-based pricing, lack of per-token rates, recent price increases, and below-average latency make it a poor choice for production workloads. Use the free credit to validate your code, then compare per-token costs with other relay stations before depositing real money.

FAQ

Q: Does the ¥4 free credit require a credit card? A: No. You get ¥4 immediately after registration with no payment method required. Q: Can I see per-token pricing before buying a package? A: No. DawCode uses package-based quotas, not per-million token rates. You cannot calculate exact cost per request without knowing the package contents. Q: What happens if I don’t use all my package quota in a day? A: The data suggests packages have daily quotas. Unused daily quota likely does not roll over, though this is not explicitly confirmed in the available information. Q: Does DawCode support batch discounts or volume pricing? A: No batch discount thresholds or volume pricing are mentioned in the platform data. You pay per package regardless of usage volume. Q: Can I get a refund if I’m unhappy with the service? A: No refund policy is specified. Assume deposits are non-refundable. Test with the free ¥4 credit first.

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.