About

island AI Coding (codex789.com) is a one-stop enterprise-grade AI API relay and aggregation gateway, providing unified access to 40+ top-tier large models including Claude Opus 5, GPT-5.6, Grok 4.6, Gemini 3.6, DeepSeek V4, Kimi K3, and Zhipu GLM. It offers an OpenAI-compatible unified Base URL and supports prompt caching, streaming, and image/video generation endpoints. The image generation API starts as low as ¥0.01 per image, and video generation costs ¥0.05 per second, focusing on stability, low prices, and multi-model access.

Key Features

Unified Access: One Base URL (https://www.codex789.com/v1) connects to 40+ models, fully compatible with the OpenAI standard API.

Prompt Caching: Natively supported, with a hit rate of up to ~96%, effectively reducing API call costs.

Image & Video APIs: Gemini 3.1 Flash Image, Grok Imagine Image 2.0/Quality (¥0.01/image), and Grok Imagine Video 1.5 (¥0.05/second), all accessible via a unified endpoint.

Streaming Output: Stable, uninterrupted streams for long sessions, with automatic traffic routing and circuit-breaker fallback.

Zero-Modification Integration: Works out of the box with Codex, Claude Code, Cursor, Cherry Studio, Chatbox, Dify, and standard SDKs.

Pricing & Fees

Pricing: Pay-as-you-go via a prepaid balance; text models billed per million tokens (input and output separately) using an "upstream price × site multiplier" formula, e.g. GPT-6.1 Sol at 0.05x ($0.10 in / $0.50 out per MT), Claude Sonnet 5.5 at 0.1x ($0.2 in / $1 out per MT).
Image and video rates: Image generation starts at ¥0.01 per successfully generated image (e.g. Grok Imagine Image 2.0); video (Grok Imagine Video 1.5) is billed at ¥0.05 per second of final output.
Currency and billing: USD-quoted models stay in $ and CNY-quoted models stay in ¥; the site does not convert currencies. Actual charges are determined by the console, which takes precedence over public page prices.
Long-input surcharge: For GPT-6 Sol / GPT-6.1 Sol, requests exceeding 272K input tokens are billed at higher rates (official input/cache at 2x, output at 1.5x), e.g. roughly $0.20/$0.75 per MT at 0.05x.
Recharge and access: Free registration to obtain an API key; top up via WeChat Pay or Alipay. Unified OpenAI-compatible Base URL (https://www.codex789.com/v1) covering 40+ models including GPT, Claude, Gemini, Grok, GLM and DeepSeek.

Who It's For

Ideal for developers, AI application teams, and enterprises that need to call large model APIs in bulk—especially teams looking to integrate image/video generation capabilities at low cost or manage multiple model channels in a unified way.

Key Advantages

Low cost & stable: Image generation from ¥0.01 per image, with multi-datacenter, multi-upstream health checks and millisecond-level intelligent scheduling.
No diluted models: Models retain their full, unquantized weights and system fingerprints, with probe-based self-verification supported.
Privacy-friendly: No persistent storage of API request bodies (prompts and model outputs).
Solves pain points: Handles multi-vendor API adaptation, account ban risk control, and concurrent stream interruption issues.

About

Key Features

Pricing & Fees

Who It's For

More

Key Advantages

Reviews (0)

Would you recommend island AI Coding? Leave a comment

Related Tutorials

Related Tools