Limited-time trial balance for new accounts

Membership-level pricing,paid by API usage

CoderPlan lets you call model APIs at membership-level cost without prepaying a monthly plan. Balance is consumed by real usage, built for fluctuating AI coding workflows across Claude Code, Cursor, Codex CLI, and Gemini CLI.

Membership-level cost, pay-as-you-go billing, with model, token, cost, cache-savings, and balance details you can inspect.

api-previewSetup guide

$ export API_BASE=https://api.coderplan.ai

$ claude "Build a Pomodoro app"

model: claude-sonnet

Input 34K · Cache read 94K · Output 4.2K

Cost ¥6.20 · Estimated savings ¥12.40

Balance ¥93.80 · Request succeeded

Membership-level pricing

API calls can reach membership-like usage cost

No full-month prepay

Avoid committing a whole month of budget upfront

Use more when busy

Pay by real task usage for fluctuating development cycles

High cache hit rate

Cache hits help the same balance cover more useful work

Auditable spend

Model, tokens, cache savings, and balance movement stay visible

Much cheaper than list price, same models

See roughly how much common models can save, then run a real task with a small balance. No monthly prepay and no budget guesswork.

Every call records model, tokens, cache savings, and balance movement so you can judge whether the spend was worth it.

GPT-5.5
OpenAI
about 0.7/10
Official price
$5.00 / $30.00
input / output
CoderPlan price
$0.35 / $2.08
input / output
GPT-5.4
OpenAI
about 0.7/10
Official price
$2.50 / $15.00
input / output
CoderPlan price
$0.17 / $1.04
input / output
Claude Opus 4.7
Anthropic
about 1.2/10
Official price
$15.00 / $75.00
input / output
CoderPlan price
$1.74 / $8.68
input / output
Claude Opus 4.6
Anthropic
about 1.2/10
Official price
$15.00 / $75.00
input / output
CoderPlan price
$1.74 / $8.68
input / output
Claude Sonnet 4.6
Anthropic
about 3.5/10
Official price
$3.00 / $15.00
input / output
CoderPlan price
$1.04 / $5.21
input / output
Gemini 2.5 Pro
Google
about 9.7/10
Official price
$1.25 / $10.00
input / output
CoderPlan price
$1.22 / $9.72
input / output

Unit: USD/1M tokens · View all model prices

The price advantage comes from cost optimization, not model replacement

CoderPlan's pricing advantage comes from model routing, cache hits, and pay-as-you-go efficiency. You are still calling Claude, OpenAI, Gemini, and other model APIs; after a call, you can see which model was used, how much cache cost was saved, and why the balance changed.

Choose a top-up tier

Top up as needed. Balance does not expire and is available anytime.

Trial

Good for first-time evaluation and one project setup

¥30
Standard balance

Daily development

Complete a small project's full development cycle

¥100
Bonus ¥5
Most popular

Deep usage

High-frequency AI coding for a week of delivery

¥300
Bonus ¥20

Professional

Heavy monthly usage and team collaboration

¥500
Bonus ¥50

New users can first use trial balance to connect a tool, run a real task, and then top up after confirming the result.

Usage path

CoderPlan focuses on three things

Easy to start

Claim trial balance, create an API Key, follow the docs for common AI coding tools, and start within minutes.

Pay as you go

Turn membership-level pricing into rechargeable balance for fluctuating API and AI coding tasks.

Know the value

See model, tokens, cache savings, and balance changes after every call, so you know which tasks deserve more budget.

Pay for outcomes

Put budget into results, not monthly lock-in

AI coding work is not evenly distributed every day. You need capacity when work is heavy, without burning monthly fees during quiet periods. CoderPlan bills real API calls so budget goes to requests that move development, debugging, and delivery forward.

Call cache optimization

The savings come from high cache hits

AI coding tools repeatedly send project context and system prompts. After a cache hit, repeated content does not need to be repriced at the full rate every time.

Cache miss92% full-billing range
Input 128KCache read 0Output 4.2KCost ¥18.60
Cache hit38% full-billing range
Input 34KCache read 94KOutput 4.2KCost ¥6.20Estimated savings ¥12.40

Example based on common long-context tasks. Actual usage changes with model and task content.

System status

Core service stability

We continuously sync key dependency status for Claude, OpenAI, Gemini, and others so you can judge whether the site is stable to use.

Status sync delayed

Updated at Waiting for sync · refreshes every 60 seconds

Last 90 days generated from upstream service status; gray means status sync is delayed.
90-day combined availability --

Claude API

Claude · Sync delayed

Sync delayed

--

90 days ago--Today

OpenAI API

OpenAI · Sync delayed

Sync delayed

--

90 days ago--Today

Gemini API

Gemini · Sync delayed

Sync delayed

--

90 days ago--Today

Trackable spend

Long-term access needs clear spend

Every request keeps model, token, cache-read, status, and balance-change details. Run small tasks first, then decide which models fit daily work and which should be reserved for hard problems.

Model and unit price
Input / output usage
Cache-read amount
This-call cost
Balance change
Call time and result

FAQ

Have another question? Contact our support team anytime.

Yes. The price advantage comes from model routing, cache optimization, and pay-as-you-go efficiency, not replacing your request with a cheaper model. Use trial balance for a real task first, then inspect quality, speed, and balance movement before continuing.

Start building your first app

Claim trial balance, connect Claude Code, Cursor, or Codex CLI, and turn an idea into a working product.