Pay-as-you-go · Pay for what you use

AI Token — one API to call leading LLMs

An OpenAI-compatible interface backed by real model routing. Send one request, reach the right model, and pay by token usage — input and output priced transparently per million tokens.

Recharge balance
Models & pricing

Real models, transparent per-token pricing

Below are the dialogue models available right now and their token rates. Prices shown are USD per million tokens for the standard tier (0–200K context).

ModelContextInputOutput
Gemini 2.5 Pro
Flagship
$0.1924per 1M tokens200K+ $0.3847$1.5386per 1M tokens200K+ $2.3077
豆包 Seed 1.6
Flagship
$0.0177per 1M tokens200K+ $0.0177$0.0440per 1M tokens200K+ $0.1759
GLM-4.7
Flagship
$0.0660per 1M tokens200K+ $0.0880$0.3077per 1M tokens200K+ $0.3517
通义千问 3-VL Plus
Flagship
$0.1429per 1M tokens200K+ $0.2143$1.4286per 1M tokens200K+ $2.1429
通义千问 3-VL Flash
Value
$0.0214per 1M tokens200K+ $0.0429$0.2143per 1M tokens200K+ $0.4286
MiniMax M2.1 极速版
Flagship
$0.0443per 1M tokens$0.1771per 1M tokens
Claude Sonnet 4.6
anthropicFlagship
200k$3per 1M tokens$15per 1M tokens
Claude Opus 4.6
anthropicFlagship
200k$5per 1M tokens$25per 1M tokens
Claude Haiku 4.5
anthropicValue
200k$1per 1M tokens$5per 1M tokens
GPT-5.4
openaiFlagship
128k$5per 1M tokens200K+ $0.7693$22.50per 1M tokens200K+ $3.4616
GPT-4o
Value
$0.3847per 1M tokens$1.5386per 1M tokens
GPT-5.1
Value
$1.25per 1M tokens$10per 1M tokens
GPT-5.2
Flagship
$0.2693per 1M tokens$2.1539per 1M tokens
Gemini 2.5 Lite
Value
$0.0463per 1M tokens$0.3847per 1M tokens
Gemini 3.0
Value
$0.0770per 1M tokens$461.5714per 1M tokens
Gemini 3.1 Lite
Value
$0.0386per 1M tokens$0.2309per 1M tokens
Gemini 3.5
Flagship
$1.50per 1M tokens$9per 1M tokens
Kimi K2.5
moonshotValue
$0.5714per 1M tokens$3per 1M tokens
Claude Opus 4.7
anthropicFlagship
200k$5per 1M tokens$25per 1M tokens
Claude Opus 4.8
anthropicFlagship
200k$5per 1M tokens$25per 1M tokens
Gemini 2.5 Pro
Flagship
Input$0.1924per 1M tokens200K+ $0.3847
Output$1.5386per 1M tokens200K+ $2.3077
豆包 Seed 1.6
Flagship
Input$0.0177per 1M tokens200K+ $0.0177
Output$0.0440per 1M tokens200K+ $0.1759
GLM-4.7
Flagship
Input$0.0660per 1M tokens200K+ $0.0880
Output$0.3077per 1M tokens200K+ $0.3517
通义千问 3-VL Plus
Flagship
Input$0.1429per 1M tokens200K+ $0.2143
Output$1.4286per 1M tokens200K+ $2.1429
通义千问 3-VL Flash
Value
Input$0.0214per 1M tokens200K+ $0.0429
Output$0.2143per 1M tokens200K+ $0.4286
MiniMax M2.1 极速版
Flagship
Input$0.0443per 1M tokens
Output$0.1771per 1M tokens
Claude Sonnet 4.6
anthropicFlagship
Context 200k
Input$3per 1M tokens
Output$15per 1M tokens
Claude Opus 4.6
anthropicFlagship
Context 200k
Input$5per 1M tokens
Output$25per 1M tokens
Claude Haiku 4.5
anthropicValue
Context 200k
Input$1per 1M tokens
Output$5per 1M tokens
GPT-5.4
openaiFlagship
Context 128k
Input$5per 1M tokens200K+ $0.7693
Output$22.50per 1M tokens200K+ $3.4616
GPT-4o
Value
Input$0.3847per 1M tokens
Output$1.5386per 1M tokens
GPT-5.1
Value
Input$1.25per 1M tokens
Output$10per 1M tokens
GPT-5.2
Flagship
Input$0.2693per 1M tokens
Output$2.1539per 1M tokens
Gemini 2.5 Lite
Value
Input$0.0463per 1M tokens
Output$0.3847per 1M tokens
Gemini 3.0
Value
Input$0.0770per 1M tokens
Output$461.5714per 1M tokens
Gemini 3.1 Lite
Value
Input$0.0386per 1M tokens
Output$0.2309per 1M tokens
Gemini 3.5
Flagship
Input$1.50per 1M tokens
Output$9per 1M tokens
Kimi K2.5
moonshotValue
Input$0.5714per 1M tokens
Output$3per 1M tokens
Claude Opus 4.7
anthropicFlagship
Context 200k
Input$5per 1M tokens
Output$25per 1M tokens
Claude Opus 4.8
anthropicFlagship
Context 200k
Input$5per 1M tokens
Output$25per 1M tokens
How billing works

Simple, usage-based billing

Pay-as-you-go

No monthly commitment. You are billed only for the tokens you actually send and receive.

OpenAI-compatible API

Use the same request format you already know. Point your existing client at our endpoint and switch models freely.

Tiered by context length

Standard rate applies to requests within 0–200K tokens; longer 200K+ requests are billed at the long-context rate where supported.

Real backend routing

Every request is routed to a real model. Input and output tokens are metered and priced separately, per million tokens.

Start calling models by the token

Download the client, sign in, and start sending requests. Pay only for what you use.

Recharge your balance