NekoCode’s “new user deal” is ¥15 for ¥30 credit, but beyond that, there’s no free trial and no published per-token pricing—you’re buying credit packages starting at a ¥20 minimum top-up.

Pricing Model and Credit System

NekoCode doesn’t charge per-million-token rates like OpenAI or Anthropic. Instead, you prepay via Alipay or WeChat Pay and consume credits based on internal conversion rates. The only concrete price anchor is the new user offer: pay ¥15, get ¥30 in credit (effectively 50% off your first ¥30 of usage). After that, standard top-ups begin at ¥20. This means cost comparison against direct API access is opaque. You can’t calculate a simple “GPT-4o input costs X per million tokens” because NekoCode doesn’t publish those rates. What you get in exchange is access to GPT-4o and Claude 3.5 Sonnet through a relay that doesn’t require a VPN—and the platform claims “MAX group quality with stable outputs.” If you’re a Chinese developer who needs reliable access more than price transparency, that tradeoff might work.

Pricing Feature Detail
New user offer ¥15 for ¥30 credit (50% off first ¥30)
Minimum top-up ¥20
Free tier None
Per-token rates Not published
Batch discounts Not available
Payment methods Alipay, WeChat Pay
Promo code None

Pros & Cons

Pros

  • New user deal cuts your first ¥30 cost in half—worth trying if you’re curious
  • MAX group claims stable outputs, which matters when you’re hitting production-grade GPT-4o or Claude 3.5 Sonnet
  • Clean interface, no bloat
  • Both MAX and Kiro groups available (though Kiro quality is unverified) Cons
  • No free tier—you must pay to test
  • Minimum ¥20 top-up locks out casual users who just want ¥5 of credit
  • No per-token pricing published, so you can’t compare against direct API costs
  • Older Claude top-ups are gone—if you wanted legacy Claude models, you’re out of luck
  • Newer station (2026 launch), so less community track record

Verdict

NekoCode is a niche play: good channel quality, minimal model selection (only two), and pricing that hides behind credit packs. If you’re a developer who values uptime and stable outputs over model variety and cost transparency, the new user deal makes it low-risk to test. But if you need to budget per-token or compare against direct API pricing, you’ll be frustrated by the lack of published rates. The ¥20 minimum top-up and no free tier mean it’s not for casual experimentation—it’s for committed users who prioritize reliability over price.

FAQ

Q: Can I calculate the per-token cost for GPT-4o on NekoCode? A: No. NekoCode doesn’t publish per-million-token rates. You buy credit in ¥20+ increments and consume based on internal conversion rates that aren’t publicly documented. Q: Is there a free trial to test the service before paying? A: No. The only discount is the new user deal (¥15 for ¥30 credit). You must pay at least ¥15 to access the platform. Q: How do I compare NekoCode’s pricing against direct API access from OpenAI or Anthropic? A: You can’t directly. NekoCode’s credit system is opaque. The value proposition isn’t cost savings—it’s VPN-free access and stable outputs. If price transparency is critical, look for a platform that publishes per-token rates.

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.