206 models, one API

One endpoint.
Every model.

OpenAI-compatible gateway to Claude, GPT, Gemini, GLM, DeepSeek, and 200+ more. Swap your base URL and everything works.

$ curl https://api.atoye.dev/v1

Get API Key Quick Start

Models

No prefix needed. Use the model name directly.

Model	Provider	Tier	Per 1M tokens
glm-5.3	Z.ai	Free	$0.00
gemini-3.8-flash	Google	Free	$0.00
deepseek-v4-flash	DeepSeek	Free	$0.00
mimo-v2.5	Xiaomi	Free	$0.00
hy3	Tencent	Free	$0.00
claude-sonnet-4.5	Anthropic	Pro	$1.50 / $7.50
claude-opus-5	Anthropic	Pro	$5.00 / $15.00
gpt-5.6	OpenAI	Pro	$2.50 / $10.00
gemini-3.7-flash	Google	Pro	$0.15 / $0.60
deepseek-v4-pro	DeepSeek	Pro	$0.27 / $1.10
grok-4.6	xAI	Pro	$2.00 / $10.00
kimi-k3	Moonshot	Pro	$0.25 / $1.00

What you get

Everything you need to ship, nothing you don't.

Drop-in compatible

Change one URL and one key. Your OpenAI SDK, LangChain, or raw HTTP calls work without code changes.

500+ models

Claude, GPT, Gemini, GLM, DeepSeek, Qwen, Grok, MiniMax. Switch by changing the model parameter.

Streaming and tools

Server-sent events for real-time output. Function calling for agents that call your code.

Free tier included

GLM 5.3, Gemini 3.8 Flash, and more at zero cost. No credit card to start.

Pricing

No monthly fee. Pay for what you use.

Free

$0/mo

For testing and production on free models.

All free-tier models
GLM 5.3, Gemini 3.8 Flash, DeepSeek V4 Flash
Streaming and function calling
60 requests per minute

Pro

Pay/token

Claude, GPT, Gemini Pro, and all premium models.

Access to all 206 models
Claude Opus 5, GPT 5.6, Grok 4.6
Higher rate limits
Below direct provider pricing

Quick Start

Three lines. No new SDK.

# curl
curl https://api.atoye.dev/v1/chat/completions \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"glm-5.3","messages":[{"role":"user","content":"Hi"}]}'

# python
from openai import OpenAI

client = OpenAI(base_url="https://api.atoye.dev/v1", api_key="YOUR_KEY")
r = client.chat.completions.create(
    model="glm-5.3",
    messages=[{"role":"user","content":"Hi"}],
)
print(r.choices[0].message.content)
  

One endpoint.Every model.