Public prices

One API for five model families
Clear pricing, pay as you go

ai2save provides a unified API relay for popular models such as GPT, Claude, DeepSeek, Kimi and GLM. Connect to development tools like Codex and Claude Code in one click, with full model quality, usage-based billing, and reliable service.

Popular model pricing

Transparent pricing · Simple choices

Featured models

USD / 1M tokens

GPT

Astra · Sol · Terra

Reference prices — not a current offer

90.0% off

Claude

Opus · Sonnet · Haiku · Fable

Reference prices — not a current offer

75.0% off

DeepSeek

Flash · V4 Pro

Reference prices — not a current offer

45.0% off

Kimi

K3.1 Pro · K3 · K2.7 Code

Reference prices — not a current offer

45.0% off

GLM

5.3 · 5.2 · 5.1

Reference prices — not a current offer

50.0% off
One API for all modelsView Model Pricing →
  • One-click setup

    CC Switch writes your configuration in one click with no manual editing. For mainstream IDE plugins and CLIs, just copy the URL and API key.

  • Full model quality

    Real users report the same model quality as the official service across a wide range of real-world tasks.

  • Pay as you go

    Every token usage record is available for inspection and can be compared individually with official token usage statistics.

  • Stable and reliable

    An account pool intelligently routes requests across multiple upstream accounts, automatically switching and retrying seamlessly when errors occur.

Transparent pricing · Simple choices

Popular model pricing

USD / 1M tokens

Public prices

USD / 1M tokens

GPT 90% off · gpt-5.5

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.50$3.00$0.05———
>272K$1.00$4.50$0.10———
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$5.00$30.00$0.50———
>272K$10.00$45.00$1.00———

The whole request is billed at the tier matching its total context (input + cache write + cache read)

GPT 90% off · gpt-5.6

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.50$3.00$0.05$0.625$0.625—
>272K$1.00$4.50$0.10$1.25$1.25—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$5.00$30.00$0.50$6.25$6.25—
>272K$10.00$45.00$1.00$12.50$12.50—

The whole request is billed at the tier matching its total context (input + cache write + cache read)

GPT 90% off · gpt-5.6-luna

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.02$0.12$0.002$0.025$0.025—
>272K$0.04$0.18$0.004$0.05$0.05—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.20$1.20$0.02$0.25$0.25—
>272K$0.40$1.80$0.04$0.50$0.50—

The whole request is billed at the tier matching its total context (input + cache write + cache read)

GPT 90% off · gpt-5.6-sol

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.50$3.00$0.05$0.625$0.625—
>272K$1.00$4.50$0.10$1.25$1.25—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$5.00$30.00$0.50$6.25$6.25—
>272K$10.00$45.00$1.00$12.50$12.50—

The whole request is billed at the tier matching its total context (input + cache write + cache read)

GPT 90% off · gpt-5.6-terra

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.20$1.20$0.02$0.25$0.25—
>272K$0.40$1.80$0.04$0.50$0.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$2.00$12.00$0.20$2.50$2.50—
>272K$4.00$18.00$0.40$5.00$5.00—

The whole request is billed at the tier matching its total context (input + cache write + cache read)

GPT 90% off · gpt-6-astra

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$1.00$5.00$0.10$1.25$1.25—
>272K$2.00$7.50$0.20$2.50$2.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$10.00$50.00$1.00$12.50$12.50—
>272K$20.00$75.00$2.00$25.00$25.00—

The whole request is billed at the tier matching its total context (input + cache write + cache read)

GPT 90% off · gpt-6-luna

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.01$0.05$0.001$0.0125$0.0125—
>272K$0.02$0.075$0.002$0.025$0.025—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.10$0.50$0.01$0.125$0.125—
>272K$0.20$0.75$0.02$0.25$0.25—

The whole request is billed at the tier matching its total context (input + cache write + cache read)

GPT 90% off · gpt-6-sol

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.20$1.00$0.02$0.25$0.25—
>272K$0.40$1.50$0.04$0.50$0.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$2.00$10.00$0.20$2.50$2.50—
>272K$4.00$15.00$0.40$5.00$5.00—

The whole request is billed at the tier matching its total context (input + cache write + cache read)

GPT 90% off · gpt-6.1-sol

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$0.20$1.00$0.01$0.25$0.25—
>272K$0.40$1.50$0.02$0.50$0.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
≤272K$2.00$10.00$0.10$2.50$2.50—
>272K$4.00$15.00$0.20$5.00$5.00—

The whole request is billed at the tier matching its total context (input + cache write + cache read)

Claude 75% off · claude-fable-5

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$2.50$12.50$0.25$3.125$5.00—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$10.00$50.00$1.00$12.50$20.00—

Claude 75% off · claude-fable-5-1

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$2.50$12.50$0.0625$3.125$5.00—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$10.00$50.00$0.25$12.50$20.00—

Claude 75% off · claude-opus-4-6

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.25$6.25$0.125$1.5625$2.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$5.00$25.00$0.50$6.25$10.00—

Claude 75% off · claude-opus-4-7

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.25$6.25$0.125$1.5625$2.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$5.00$25.00$0.50$6.25$10.00—

Claude 75% off · claude-opus-4-8

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.25$6.25$0.125$1.5625$2.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$5.00$25.00$0.50$6.25$10.00—

Claude 75% off · claude-opus-5

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.25$6.25$0.125$1.5625$2.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$5.00$25.00$0.50$6.25$10.00—

Claude 75% off · claude-opus-5-5

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.00$5.00$0.05$1.25$2.00—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$4.00$20.00$0.20$5.00$8.00—

Claude 75% off · claude-sonnet-4-6

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.75$3.75$0.075$0.9375$1.50—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$3.00$15.00$0.30$3.75$6.00—

Claude 75% off · claude-sonnet-5

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.50$2.50$0.05$0.625$1.00—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$2.00$10.00$0.20$2.50$4.00—

Zhipu GLM 50% off · glm-5.1

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.616$1.936$0.1232$0.44$0.44—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.40$4.40$0.28$1.00——

Zhipu GLM 50% off · glm-5.2

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.5188602762$1.816010967$0.1297150691$0.5188602762$0.5188602762—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.179227901$4.127297652$0.2948069751$1.179227901——

Zhipu GLM 50% off · glm-5.3

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.616$1.936$0.1144———
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.40$4.40$0.26———

Kimi 45% off · kimi-k2.7-code

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.5269674681$2.18894179$0.1053934936$0.5269674681$0.5269674681—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.9581226692$3.979894164$0.1916245338$0.9581226692——

Kimi 45% off · kimi-k3

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.621438363$8.107191816$0.1621438363$1.621438363$1.621438363—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$2.948069751$14.74034876$0.2948069751$2.948069751——

Kimi 45% off · kimi-k3.1-pro

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$1.10$5.50$0.11$1.10$2.20—
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$2.00$10.00$0.20$2.00$4.00—

Deepseek 45% off · deepseek-flash

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.08249992$0.32999968$0.00165026———
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.15$0.60$0.003———

Time-based pricing (optional) · UTC · Weekdays

01:00–04:00 ×2

06:00–10:00 ×2

Deepseek 45% off · deepseek-v4-flash

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.08249992$0.32999968$0.00165026———
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.15$0.60$0.003———

Time-based pricing (optional) · UTC · Weekdays

01:00–04:00 ×2

06:00–10:00 ×2

Deepseek 45% off · deepseek-v4-pro

token

Site Price (Discounted)
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.36300052$1.08900047$0.01210009———
Official Price
Context (tokens)InputOutputCache readCache write (5m)Cache Write (1h)USD / request or image
Standard$0.66$1.98$0.022———

Time-based pricing (optional) · UTC · Weekdays

01:00–04:00 ×2

06:00–10:00 ×2

Standard-time prices. Context tiers, cache duration, time multipliers and per-request units are listed below. The selected group’s billing rules apply.

Public prices →

Retrieved:

Source: OpenAI · Anthropic · Gemini · LiteLLM — Not independently verified

Understand it first

What this relay is for

It serves users who want affordable access to frontier models, with transparent pricing, verification, privacy, and refund rules.

Why the price is lower

A relay brings multiple upstream accounts and model entry points behind one gateway, then bills by actual usage. Lower pricing comes from pooled purchasing, shared quota, and automatic scheduling; upstream model costs and official limits still apply.

  • Multiple users share subscription quota and split fixed monthly, credit, and operating costs.
  • Official discounts, credits, and team benefits are used, so the price should remain traceable to upstream costs.
  • The account pool routes requests by model, region, and live quota to reduce idle capacity and waste on a single account.
  • Pay for actual text, image, or other usage instead of buying a full month of capacity you may not use.

Who it is for

Individual developers, students, independent creators, and teams that need to try several models for a short period. It fits prototypes, coding assistance, drafts, and model comparisons; evaluate official plans first when you require direct access, enterprise compliance, or highly confidential data processing.

How to detect model substitution

Prepare repeatable fixed tasks before buying and run them with the same model, prompt, and parameters. Compare structured output, long-context handling, code repair, and image understanding; then check the model name, latency, error codes, and billing records. One polished demo or a low price alone is not evidence.

Parrot SVG test (full prompt)

Output only one complete SVG file that can be rendered directly. Do not use Markdown code fences and do not add any explanation.

Canvas: width=800, height=600, viewBox="0 0 800 600", with a transparent background. Draw a cartoon parrot facing left and centered on the canvas. Use green for the head, yellow for the cheek, orange for the beak, blue for the chest, and red plus purple for the wing, with three feather layers to show depth.

Details: add a clear eye with a highlight, claws, tail feathers, and a branch. Use 8px rounded strokes, smooth edges, and sufficient color contrast. Keep every shape inside valid SVG, g, path, circle, or ellipse elements. Do not reference external images, fonts, CSS, JavaScript, or data URIs.

Self-check: close every XML tag, make every path d attribute valid, keep the artwork inside the viewBox, and include title and desc elements in the SVG. After the SVG, output nothing else.

Verify: 1) it renders directly; 2) the colors, direction, and composition match; 3) the source has readable groups, layers, and paths; 4) the same prompt produces stable results through the official channel and the relay.

During a full official outage

A relay depends on official upstream services. When official servers are broadly unavailable, a normal relay cannot continue either. Check the official status page first, then send a short request with the same model and prompt. Record the HTTP status, model name in the response, latency, and billing record. If the official service is broadly down but a platform claims that every model is healthy, it may be switching models, using cache, or using an undeclared channel. Pause top-ups and ask for an explanation.

Privacy and risks

By design, a relay can potentially see requests, context, and replies. Treat the content as sensitive.

  • Do not hand a relay, ChatGPT, or a coding assistant your full account, email, cookies, or computer control. Grant only the minimum files and permissions required by the current workspace.
  • Remove API keys, passwords, identity documents, customer data, and business secrets before sending. Replace them with placeholders and instruct the model not to reconstruct or print those fields.

Account pool advantages

  • Typical path: unified gateway → authentication and quota check → model routing → available upstream account → streamed response.
  • If an account is rate-limited, disconnected, or out of balance, the scheduler can retry and switch to another account for the same model, reducing single-account failures.
  • The pool can also route by region, concurrency, and remaining quota; it cannot bypass an official ban, model retirement, or a broad upstream outage.

How to use it

Follow the documentation to get an API key, configure your client, and verify calls step by step. Start with a low-balance key and a simple model to confirm the endpoint, auth header, and model name before connecting an IDE, script, or production service. Monitor usage, error codes, and balance regularly. Keep the request time and response ID when something fails so upstream and routing issues can be traced.

Open the guide

Why choose ai2save

Turn complex AI subscriptions into one simple choice

Find popular discounts

We collect worthwhile AI deals every day so you can compare prices first.

One entry for many models

Use one API for GPT, Claude and other leading models without repeated setup.

Pay as you use

Clear usage billing keeps both budgets and consumption under control.

Integration docs

From setup to first request

Choose a client guide to prepare your environment, configure access, and connect to the API.

View all docs

Find your next AI model today

Join ai2save to try the latest models at a lower cost.

Get Started

© 2026 ai2save · Make every AI call more worthwhile