OpenAI-compatible inference

Inference without refusal.

Run leading open models through the OpenAI SDK with long context, predictable pricing, and no provider lock-in.

Four models. One endpoint. No usage tiers.

PythonQuickstart
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.openai.com/v1""https://inference.coret.ai/v1",
)

response = client.chat.completions.create(
    model="gpt-4o""coret/deepseek-v4-flash",
    messages=[{"role": "user", "content": "Hello"}],
)
Connected·OpenAI-compatible·Streaming ready

Four models. One endpoint.

Choose for speed, depth, or scale. Switch with a single model name.

Qwen3.8 27B Uncensored

262K ctx
coret/qwen3.8-27b-uncensored

No refusal training. Built for security research, fiction, and blunt subject matter.

Input
$0.20
Cached input
$0.05
Output
$0.80

DeepSeek V4 Flash

1M ctx
coret/deepseek-v4-flash

Million-token context reads whole repos, logs, and archives in one pass.

Input
$0.054
Cached input
$0.011
Output
$0.108

GLM 5.2

256K ctx
coret/glm-5.2

Multi-step reasoning and dependable tool calls for work that gets reviewed.

Input
$0.80
Cached input
$0.22
Output
$1.80

All prices USD per 1M tokens.

Switch in two strings.

Keep your OpenAI SDK. Change the base URL and model name.

migration.py
2 changes
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://inference.coret.ai/v1",
)

response = client.chat.completions.create(
    model="coret/qwen3.8-27b-uncensored",
    messages=["role": "user", "content": "Hello"}],
)

Everything else stays the same.

Read the API docs

Your first request is minutes away.

Get a key through AntSeed, point your OpenAI client to Coret, and start building.

  1. Create account
  2. Add Coret provider
  3. Send request

Provider “Coret” · agent #60566 · antseed.com