The best way to evaluate Hober is to use it. The free tier gives every developer key a daily allowance of requests against whichever models Hober's catalog currently flags free, enough to build a prototype, test an integration, or explore the API. Create an API key and start making requests. No credit card and no deposit required for a standard developer key. Your current allowance and how much of it you have used are returned in the X-FreeQuota-Limit and X-FreeQuota-Used headers on every free-model response, with X-FreeQuota-Remaining added while you are still under the cap, and the daily cap lifts entirely once you move to a paid tier.

Free models are not a pass-through to some upstream provider's own free tier. Hober still pays the real per-token cost to the upstream provider on every free-model call. It just does not charge you for it. Before billing runs, the gateway checks whether the requested model carries a free flag in its catalog record, and skips the balance check when it finds one. Exactly one model carries that flag today: glm/glm-4.5-flash. Query GET /v1/models for the live list instead of trusting a hand-typed table here. Each entry in that response reports its own free status, read from the same catalog record the billing middleware checks, so the two can never disagree. When you're ready for premium models like GPT-4o or Claude, upgrade by depositing credits or connecting a wallet for x402 payments.

Limits

Eligible Models

The set of free-eligible models changes as Hober's catalog changes, so this page does not hardcode a list that would go stale. Call GET /v1/models and filter the response for entries where free is true. That field is read from the same catalog record the billing middleware checks, so it cannot drift out of sync with what actually gets billed. One model carries that flag as of this writing: glm/glm-4.5-flash.

Usage

// glm/glm-4.5-flash is free-eligible today; confirm anytime via GET /v1/models
{
  "model": "glm/glm-4.5-flash",
  "messages": [{"role": "user", "content": "Hello"}]
}

// Or force the free-tier path explicitly with the :free suffix
{
  "model": "glm/glm-4.5-flash:free",
  "messages": [{"role": "user", "content": "Hello"}]
}