Use Claude Code with GLM 5.2, Kimi K2.7 or DeepSeek

Claude Code works with CheapestInference out of the box: the API exposes an Anthropic-compatible Messages endpoint, so you only need to set two environment variables and pick a model. During your reserved time blocks, usage is unlimited — no per-token billing.

Setup (2 minutes)

1. Subscribe to a pool and copy your subscriber key (sk-...) from the dashboard.
2. Export these variables in your shell (or put them in your shell profile):

export ANTHROPIC_BASE_URL="https://api.cheapestinference.com/anthropic"
export ANTHROPIC_AUTH_TOKEN="sk-..."   # your subscriber key
export ANTHROPIC_MODEL="glm-5.2"       # any model in your pool
export ANTHROPIC_SMALL_FAST_MODEL="glm-5.2"

3. Start claude as usual. All requests now run against your pool.

Which model id to use

You want	Set ANTHROPIC_MODEL to	Pool
GLM 5.2	glm-5.2	Frontier
Kimi K2.7	kimi-k2.7	Frontier
Kimi K2.6	kimi-k2.6	Frontier
MiniMax M3 (1M context)	MiniMax-M3	Frontier
DeepSeek V4 Flash (1M context)	deepseek-v4-flash	Core
MiMo v2.5 (1M context)	mimo-v2.5	Core

A subscription covers every model in its pool, so you can switch models by changing the variable — no new key, no new plan.

Per-project configuration

To pin a project to a specific model without touching your global shell profile, create .claude/settings.json in the project:

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.cheapestinference.com/anthropic",
    "ANTHROPIC_AUTH_TOKEN": "sk-...",
    "ANTHROPIC_MODEL": "kimi-k2.7",
    "ANTHROPIC_SMALL_FAST_MODEL": "kimi-k2.7"
  }
}

Notes

• Unlimited during your blocks: within your reserved time blocks there are no token caps. Outside your blocks, requests are rejected until the next block starts — see Unlimited Subscriptions.
• Thinking: models that emit reasoning surface it through the standard Anthropic thinking content blocks, so Claude Code renders it natively.
• Other coding agents: the same two variables work for any tool that speaks the Anthropic API. Tools that speak the OpenAI API instead (Cline, Aider, OpenCode, Cursor with a custom endpoint) should use https://api.cheapestinference.com/v1 — see Chat Completions.