Unified model API Multi-model access, transparent pricing

A professional and reliable large-model API provider One key for multiple models

Monthly plans cover multiple models, with rollover quota, stackable packages, high cache hits, and flexible pay-as-you-go.

View Packages

Connected providers

Works with your existing tools

Connect mainstream coding tools in minutes.

Models

LLM / image / video / audio coverage

Providers

Route multiple providers with one key

Public packages

Mix packages and pay-as-you-go

Featured Models

Featured models

View all models

GPT-5.3 Codex Spark

OpenAI / gpt-5-3-codex-spark

Lightweight Codex-oriented GPT model for quick code Q&A, small edits, and lower-cost development tasks.

GPT-5.5

OpenAI / gpt-5-5

OpenAI flagship model for complex reasoning, coding, architecture design, and demanding tasks.

Claude Opus 4.8

Anthropic / claude-opus-4-8

Claude high-end reasoning model for complex coding, architecture, and analysis.

Claude Sonnet 4.6

Anthropic / claude-sonnet-4-6

Balanced Claude model for speed and intelligence across development, writing, and complex Q&A.

Claude Haiku 4.5

Anthropic / claude-haiku-4-5

Fast lightweight Claude model for low-latency, high-concurrency, and cost-sensitive tasks.

GPT-5.4 Mini

OpenAI / gpt-5-4-mini

OpenAI lightweight model balancing speed and cost for daily Q&A, batch jobs, and low-latency use cases.

GPT-5.4

OpenAI / gpt-5-4

OpenAI high-performance general model for coding, professional writing, and complex business workflows.

Tool setup

Quick setup for multiple agents.

Ready-made templates for common agents. Copy them into the right files, add your SU8 API key, and start calling models.

One gateway for API keys and base_url settings
Ready-to-copy config for each agent, with fewer fields to edit
Request logs, package billing, and balance usage stay easy to inspect

View Docs

Config file

~/.codex/config.toml

model_provider = "su8"
model = "gpt-5.5"
network_access = "enabled"
disable_response_storage = true
model_verbosity = "high"

[model_providers.su8]
name = "su8"
base_url = "https://www.su8.codes/v1"
wire_api = "responses"
requires_openai_auth = true

Config file

~/.codex/auth.json

{
  "OPENAI_API_KEY": "su8-..."
}

Config file

~/.config/opencode/config.json

{
  "npm": "@ai-sdk/openai-compatible",
  "options": {
    "baseURL": "https://www.su8.codes/v1",
    "apiKey": "su8-..."
  },
  "models": {
    "gpt-5.5": {
      "name": "gpt-5.5",
      "variants": {
        "low": {
          "reasoningEffort": "low"
        },
        "medium": {
          "reasoningEffort": "medium"
        },
        "high": {
          "reasoningEffort": "high"
        },
        "xhigh": {
          "reasoningEffort": "xhigh"
        }
      }
    }
  }
}

Config file

~/.claude/settings.json

{
  "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1",
  "DISABLE_ERROR_REPORTING": "1",
  "DISABLE_TELEMETRY": "1",
  "env": {
    "ANTHROPIC_AUTH_TOKEN": "su8-...",
    "ANTHROPIC_BASE_URL": "https://www.su8.codes",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "gpt-5.5",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "gpt-5.5",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "gpt-5.5",
    "ANTHROPIC_MODEL": "gpt-5.5",
    "CLAUDE_BASH_NO_LOGIN": "1"
  },
  "includeCoAuthoredBy": false,
  "model": "gpt-5.5",
  "attribution": {
    "commit": "",
    "pr": ""
  }
}

Config file

~/.openclaw/openclaw.json

{
  "baseUrl": "https://www.su8.codes/v1",
  "apiKey": "su8-...",
  "api": "openai-completions",
  "models": [
    {
      "id": "gpt-5.5",
      "name": "gpt-5.5"
    }
  ]
}

Config file

~/.hermes/config.yaml

model:
  default: gpt-5.5
  provider: custom:su8
  max_tokens: 128000
custom_providers:
  - name: SU8
    key_env: SU8_API_KEY
    base_url: https://www.su8.codes/v1
    api_mode: codex_responses
    models:
      gpt-5.5:
        context_length: 256000

Config file

~/.hermes/.env

SU8_API_KEY=su8-...

Recommended Plans

Subscription packages for steady usage

Unused quota can carry over. Packages and pay-as-you-go balance can coexist on the same account.

View full pricing

Unused quota can carry over. Packages and pay-as-you-go balance can coexist on the same account.

View full pricing

Pay-as-you-go

Best for trials, light usage, and burst demand. Top up balance first, then usage is charged by the actual model calls.

All enabled models support paygo

Realtime logs and billing details

Works with packages; balance covers overflow

Community support

View paygo pricing

FAQ

FAQ.

What is SU8?

SU8 lets you call OpenAI, Claude, Gemini, and other models with one API key. Pricing is public, billing supports packages and pay-as-you-go, and calls use a unified routing and usage record flow.

Which models and providers are supported?

SU8 currently covers mainstream modern models from OpenAI, Anthropic, Google, Mistral, Alibaba, ByteDance, and more. The model page shows the current catalog.

How does billing work?

Packages are better for steady usage. Pay-as-you-go balance is better for trials or bursts. Both can coexist on the same account.

Is it compatible with OpenAI SDKs?

Yes. Point the OpenAI SDK base_url to SU8 and reuse your existing Chat Completions code. Additional protocol adapters are being expanded.

Does the console show usage?

Yes. The console includes request logs, token usage, cost analysis, key management, package quota, and per-key troubleshooting.

Ready to build?

Registration is free. Pricing and models are transparent, so you can evaluate before committing.

Paygo pricing