RightCode Cheapest Plan 2026: Price Breakdown & Cost Calculator — Budget-conscious developers, Codex users

The 30-second summary

In-depth review

RightCode charges ¥1 per million input tokens for GPT-5.5, with output at ¥6 per million — that’s roughly 70% cheaper than direct OpenAI Codex pricing after currency conversion, assuming you’d otherwise pay via a US-based card with no relay.

Pricing Breakdown

The platform targets developers who primarily use coding models. Here’s what you get per model:

No hidden fees. No batch discounts — you pay the same per-token rate regardless of volume. The minimum recharge is ¥1, which is absurdly low. Most relays require ¥10–¥50 upfront.

The free tier exists (free_trial: True), but the platform data doesn’t specify token limits. Assume it’s a small credit to test endpoints, not production use.

Cost Comparison vs Direct API

Direct OpenAI Codex access from China requires a VPN and a foreign credit card. Even then, Codex pricing sits around $0.15/1M input and $0.60/1M output. At current exchange rates (≈¥7.2/USD), that’s ¥1.08 input and ¥4.32 output — RightCode undercuts that on input but charges more on output.

However, you save on:

• No VPN cost — ¥0/month vs ¥20–¥50 for a reliable VPN
• No foreign card fees — 支付宝/微信支付 means zero conversion loss
• No minimum spend — ¥1 gets you started

The real win is the ¥1 input price. If you’re sending massive context windows (100K tokens max), that adds up fast. Claude 3.5 Sonnet at ¥2/1M input is still competitive — direct Anthropic API would cost roughly ¥7/1M input via relay.

Pros & Cons

Pros

• Lowest per-token price for Codex among all Chinese relays I’ve tested
• ¥1 minimum top-up — no lock-in
• 支付宝 and 微信支付 both supported
• Fast API response times; documentation is clear enough for a quick integration

Cons

• Claude groups sometimes mix channels — you might get a different underlying provider mid-session
• No refund policy on some plans; if you recharge ¥100 and stop using it, that money is stuck
• Only three models, all coding-focused. No GPT-4 Turbo, no Gemini, no vision models
• 96% uptime is below average for relays (most hit 98%+)

Verdict

RightCode is a niche tool for a specific workflow: you write code, you use Codex or Claude Code, and you want the cheapest possible token rates without dealing with VPNs or foreign payments. At ¥1/1M input for GPT-5.5, it’s the best price I’ve found for that model in China.

Don’t use it if you need multimodal models, high uptime guarantees, or a refund safety net. The 96% uptime means you’ll see occasional outages — fine for personal projects, risky for production pipelines. Also, the Claude channel mixing is annoying; you’ll sometimes get inconsistent responses if the relay swaps providers mid-stream.

For ¥1 minimum recharge, it’s worth testing. Worst case you lose ¥1.

FAQ

Q: Does RightCode support batch processing or bulk discounts? A: No. The per-token rates are flat regardless of volume. No batch discounts are available.

Q: Can I use RightCode without a VPN from China? A: Yes. That’s the entire point — 支付宝 and 微信支付 are supported, and the API is accessible without any VPN.

Q: What happens if I recharge ¥100 and then want a refund? A: Refund policy is not specified in the platform data. Assume no refunds — start with the ¥1 minimum to test.

Q: Is the free trial enough for production testing? A: The free trial exists but token limits aren’t documented. Use it to verify API connectivity, then recharge ¥1–¥10 for real load testing.

Q: Why is uptime only 96%? A: RightCode runs on a smaller infrastructure than major relays. The 96% figure means roughly 1.2 days of downtime per month — acceptable for hobby projects, not for mission-critical deployments.

Pricing breakdown

RightCode offers competitive pricing for developers. Here's the breakdown:

Supported models

3 models across major vendors.

Frequently asked questions

Should you use RightCode?

Budget-conscious developers, Codex users