Free trial

Free LLM API Trial: Get 1M Tokens

Test GPT, Claude, DeepSeek and Qwen-compatible models with one OpenAI-compatible API.

Buy an API keyDocsAdd usage credits
GPT-compatibleClaude-compatibleGemini-compatibleDeepSeekQwen / GLM / KimiStep / MiMo / MiniMax
  • 1M trial tokens on email signup
  • OpenAI-compatible API, no credit card
  • Multi-model gateway — not an official vendor
  • Valid for 7 days, first-time signups only
Trial quota1M tokens
Valid until7 days

Enter your email to claim

We send a verification code first to confirm your email.

Trial signups opening soon

Free trial slots are being prepared. You can buy a paid balance now and start right away.
Buy an API keyDocs

Start testing in three steps

1

Verify your email

Enter your email and the code we send you.

2

Get your API key

A trial key is created automatically and shown here and by email.

3

Start testing

Connect with curl, the OpenAI SDK, or Claude Code.

Model routes you can test

We are an OpenRouter-style multi-model gateway aggregator with compatible access to leading global and Chinese models.

GPT-compatible

Test GPT-family compatible models through one gateway.

Claude-compatible

Supports Claude Code and the Messages-compatible protocol.

Gemini-compatible

Connect to Gemini-family compatible routes.

DeepSeek

Low-cost models — trial quota goes further.

Qwen / GLM / Kimi

Covers Qwen, Zhipu GLM, Kimi and more Chinese models.

Step / MiMo / MiniMax

StepFun, Xiaomi MiMo, MiniMax and others, added over time.

We are a multi-model gateway aggregator providing compatible access. We are not OpenAI, Anthropic, DeepSeek or Qwen. Model availability, rate limits and context length vary by plan and stock.

Code samples

Set the Base URL and API key, then call it like any OpenAI-compatible endpoint.
curl https://gpt-agent.cc/v1/chat/completions \
  -H "Authorization: Bearer $YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Quota and rules

1M LLM Agent trial tokens, valid for 7 days.

Models are metered by multiplier: low-cost models like DeepSeek and Qwen go further, while GPT and Claude consume faster and have separate rate limits.

For development testing only — not for bulk production calls.

First-time signups only. Reselling, account farming and automated abuse are prohibited.

Claims from datacenter IPs, VPNs or proxies are routed to manual review.

FAQ

Is a credit card required?

No. Verify your email and the trial quota is granted.

Which models are included?

GPT, Claude and Gemini-compatible routes, plus DeepSeek, Qwen, GLM, Kimi, Step, MiMo and MiniMax. See the page and docs for the current list.

Can I use Claude Code?

Yes. Set ANTHROPIC_BASE_URL and ANTHROPIC_AUTH_TOKEN to connect.

How is 1M token calculated?

It is a unified LLM Agent trial quota; each model draws down at its own multiplier.

What happens after free tokens are used?

Upgrade to a paid balance on the buy or top-up page. Quota does not expire mid-cycle.