01
Buy the New Plan
Go to the plans page, pick a higher-quota plan and complete the payment.
Multi-Model · One API
Choose Your Plan — Multiple Model Families, One Quota
One-click presets or custom quota — one quota works across Claude, GPT, and Chinese models GLM, Kimi, DeepSeek, Qwen
- Instant Activation
- Custom Config
- Multi-Month Discount
- CLI · Desktop · IDE ready
Preset plans or a custom quota
Pick a preset plan for quick start, or customize your configuration with real-time pricing
Not sure how much quota you need? Start with Starter
Start with the Starter plan ($8.57/mo) and upgrade to a higher-quota plan whenever you're ready. The unused time on your old plan is credited back to your QCode balance pro rata, so nothing is wasted — just email hi@qcode.cc when you upgrade, and we handle it within 1 business day.
1
Daily Spend Limit
Set daily API spend cap (USD)
$5$500
2
Subscription Duration
Subscribe for multiple months to enjoy discounts
≈ tokens (reference)
Sonnet 5
Opus 5 / 4.8
Fable 5
Haiku 4.5
Includes caching; actual usage varies by model mix
Configuration Info
Daily Limit
Max daily API spend, resets at 00:00 Beijing time
Multi-Month Discount
Discounts from 3 months: 3mo 10% off, 6mo 20% off, 12mo 25% off
Plan Comparison
Compare all 7 preset plans at a glance
All plans below support Claude, GPT and Chinese models GLM, Kimi, DeepSeek, Qwen — one quota across every platform
| Feature | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Daily Limit | ||||||||||
| Concurrency (per Key) | 5 | |||||||||
| Request rate | Unlimited | |||||||||
| Unit Price | ||||||||||
| Models | All Claude + GPT + 8 CN models (GLM / Kimi / DeepSeek / Qwen) | |||||||||
| View all model pricing → | Billing documentation → | ||||||||||
| Codex SupportGPT-5.6 | ||||||||||
| Chinese models work via Claude Code or the OpenAI-compatible endpoint; Codex (Responses protocol) currently supports GPT models only. | ||||||||||
| Select Plan | ||||||||||
Compare all 7 plans
Daily Limit:
Concurrency (per Key): 5
Request rate: Unlimited
Models: Multiple families (incl. 8 CN models)
Chinese models work via Claude Code or the OpenAI-compatible endpoint; Codex (Responses protocol) currently supports GPT models only.
All plans: no 4-hour window (fully retired July 2026) · 30-day validity · technical support included
// THE GAP AT A GLANCE
Same $1 of official quota — how big is the gap?
Billed per token; per-model rates are listed on /models — the only difference is what you pay per $1 of quota
Official direct
$1.00/$1 quota
save 95% · ≈20× cheaper
QCode base
$0.057
Your price（Basic plan）
$0.049
Official unit prices per LiteLLM list-price snapshot (2026-08-06). Linear scale; the two left markers are staggered for legibility.
// VALUE CONVERTER
How much quota does your budget buy?
Drag the slider — feel the gap
$70
$15$700
Estimated at the Basic plan rate ≈$0.049/$1; actual plans range ≈$0.046–0.060/$1
Quota you get
$1,429
Same quota, official price
$1,429
→
You save
$1,358.57 (−95.1%)
≈ Tokens you can burnSonnet 5, cache-inclusive estimate
≈ 540M tokens
≈ Equivalent to heavy use forTotal-volume basis at $5/day (quota resets per plan cycle; estimate)
≈ 292
Client-side estimate, not a binding quote.
What is 'Quota'?
$1 quota = $1 of API spend。You purchase a daily cap on API spend
Billed per token
Per-model rates are listed on /models
vs Pay-As-You-Go
PAYG proxies: 1.2x+ markup, price varies；QCode Basic: $51.43 → $35/day
Quick lookup
| Budget | Quota | Official value | You save |
|---|---|---|---|
| $8.57 Starter | $150 | $150 | $141.43 |
| $51.43 Basic | $1,050 | $1,050 | $998.57 |
| $117.14 Professional | $2,550 | $2,550 | $2,432.86 |
| $241.14 Ultimate | $4,000 | $4,000 | $3,758.86 |
Client-side estimate, not a binding quote.
Detailed Quota Limits Explanation
Daily Limit
The maximum API spend per day, reset at 00:00 Beijing time. If you run out during the day, you can instantly top up today's quota with a self-service Booster Pack. Booster quota resets separately at 02:00 Beijing time the next day — deliberately 2 hours later than the daily limit, leaving a buffer for work that runs past midnight.
Suggestion: Allocate usage reasonably, avoid concentrated high consumption
Rate Window (Removed)
As of July 2026, personal plans have no 4-hour window at all — no per-window spend or request caps. Use your daily quota at any time.
The daily limit is now the only usage cap
Important: these limits keep the service stable and fair for everyone. Normal development rarely hits them — contact support for special needs.
Guarantees & Upgrades
If an account gets banned, we replace it
If the Anthropic Claude account used by the subscription service gets banned, we will promptly replace it with a new working account to restore your service. You don't need to do anything - our technical team will handle the entire switching process.
Note: This guarantee covers account bans. For Anthropic usage limits (such as temporary cooldowns), these are normal platform policies and will recover automatically.
Timely Issue Detection
Through user feedback and our daily checks, we can quickly identify account anomalies and immediately initiate emergency handling procedures.
Quick Response Handling
Personal plans: usually respond within 1-2 hours during service hours (10:00–22:00), within 24 hours outside them, to restore your service as soon as possible.
For You: Completely Seamless
Just focus on development — no need to worry about technical details. Our team quietly keeps the service stable in the background.
Instant Activation
Fully automated instant activation
One key
Call Claude, GPT and leading Chinese models
Professional Technical Support
With years of Claude service operation experience and professional capabilities to handle various account issues, providing reliable technical support.
Flexible upgrade
The new plan takes effect immediately after purchase (fully self-service); to close the old plan, email hi@qcode.cc or contact online chat — the remaining value is prorated back to your balance, usually within 1 day of receiving your request (we check email on weekends too)
02
Email Your Request
After purchase, email hi@qcode.cc (or message live support in the bottom right) to request closing your old lower-quota plan.
03
Refunded to Your Balance
We usually process within 1 day of receiving your email (we check email on weekends too). The remaining value of your old plan is pro-rated and refunded to your QCode balance, usable for renewals or any spending.
Example
Say you bought Basic at $51.43/mo ($35/day) and upgrade to Standard at $117.14/mo ($85/day) after 15 days: buy Standard as usual, then request to close Basic — the remaining ~15 days (≈ $25.7) are prorated back to your balance.
// WHY SO CHEAP
Why so cheap? The cost structure
Three verifiable facts — not marketing copy
No diluted quota: billed per token
Per-model rates are listed on /models, and every charge is itemized in your dashboard. Cheap doesn't mean watered down.
Cache & scheduling dividend
Prompt-cache hits cost less, and smart scheduling raises utilization — the engineering savings show up directly in your rate. (Here "cache" means the model's official prompt cache; QCode's relay layer stores none of your request data.)
Subscriptions + daily caps shave peaks
Prepaid plans make usage predictable and daily limits smooth peak load — higher utilization, less waste, lower unit price.
Official-price comparisons on this page use the LiteLLM list-price snapshot (2026-08-06).
Frequently Asked Questions
Common questions about pricing and services
MOST ASKED
Which platforms does the plan cover? Is it Claude Code only?
Not just Claude Code. Every plan covers multiple model families with the same quota: Anthropic Claude (Opus 5 / Opus 4.8 / Sonnet 5 / Fable 5), OpenAI GPT (GPT-5.6 Sol / Terra), plus Chinese models GLM, Kimi, DeepSeek and Qwen. One key over a standard API — Claude Code, Codex CLI and other CLIs, desktop apps and IDE plugins alike; any tool that allows a third-party API endpoint can plug in. Chinese models are called via Claude Code (Anthropic protocol) or the OpenAI-compatible endpoint — Codex's Responses protocol does not support Chinese models yet.
MOST ASKED
What if I outgrow this plan? Can I upgrade and pay only the difference?
Yes. Buy a higher plan first, then email hi@qcode.cc or message live support to close the old plan — the remaining time is pro-rated back to your balance, usually processed within 1 day of receiving your request (we check email on weekends too), so you can safely start small.
Look at your monthly API spend: $150–$300/month → Basic ($35/day); $300–$500/month → Regular ($50/day); $500+/month → Standard ($85/day). Monthly plans beat pay-as-you-go in two ways: ① lower effective rate ($0.05-0.06 per $1 quota across tiers); ② booster packs available anytime for instant quota expansion. Unsure how much you need? Start with the Starter plan ($8.57/month) and upgrade once you know your usage.
We offer subscription plans, each valid for 30 days, with Alipay and other major payment methods. Fully self-service: top up, activate instantly, and start using right away — no manual processing.
QCode offers Anthropic- and OpenAI-compatible endpoints, so one API key is all you need; what differs is how you pay: flexible subscription pricing, localized support, and more payment options.
Yes. To upgrade, just buy a higher-quota plan at any time, then email hi@qcode.cc to close the old one — its unused time is credited back to your QCode balance pro rata. Each plan has its own 30-day validity and several plans can run at once. To downgrade, simply buy the lower tier after your current plan expires.
We provide technical support to all users. After purchasing an API Key, you can reach us anytime via the on-site online chat, or email us at hi@qcode.cc — we reply as soon as possible during working hours.
QCode is an API service billed per token: you connect with a QCode API key, with no separate official account to register or manage. Per-model rates and the models you can call are listed on /models.
Three verifiable facts: ① billed per token, with per-model rates listed on /models — every charge is reconcilable line by line in the dashboard; ② prompt-cache hits cost less, and smart scheduling raises utilization — the savings show up directly in our rates; ③ subscriptions + daily quotas smooth the peaks, so utilization stays high with little waste. Cheap because of these three — not watered-down quota.
We sell digital goods delivered online — once activated, resources are allocated immediately and billing begins, so we do not offer refunds. The one exception is upgrading: after buying a higher-quota plan, email hi@qcode.cc and the unused time on the old plan is credited back to your QCode balance pro rata (to your balance, not to the original payment method). In addition, we support balance transfers between QCode members: you can sell your balance on platforms such as Xianyu or Zhuanzhuan, and once you find a buyer, email hi@qcode.cc to request a transfer (please include the transfer amount and the recipient's email address). We process it within 1 business day. Note: transfers between members only — cash withdrawals are not supported.
Ready to Start Your Development Journey?
Register now and enjoy Claude Code services with flexible subscription pricing