FreeInference Documentation

Free LLM inference for coding agents and IDEs

FreeInference provides free access to state-of-the-art language models for coding agents and AI-powered IDEs, with a particularly smooth setup path for Kilo Code.

Key Features

Free Access

Free inference for coding agents and development tools

Multiple Models

Access GLM, Qwen, MiniMax, and other powerful models

IDE Integration

Easy setup with Kilo Code, Cursor, Roo Code, and more

Cloud Agents (Beta)

Delegate a task, get back a draft pull request — agents run in isolated cloud sandboxes with your choice of agent and model

Kilo-Friendly Setup

Detailed Kilo Code instructions for a fast OpenAI-compatible configuration

Getting Started

  1. Get your API key - Register at https://freeinference.org and create your API key

  2. Choose your IDE:

    • Kilo Code - Kilo Code setup with recommended models

    • Cursor - AI-powered code editor

  3. Configure and start coding!

See the Quick Start guide for detailed setup instructions.

Available Models

Model

Access

Context Length

Best For

GLM-5.3 Flash recommended

Free

1M tokens

General coding and agentic work; always reasons

DeepSeek V4 Flash agentic

Free

1M tokens

Complex coding and long reasoning chains

Qwen3.6 35B fastest

Free

262K tokens

Quick edits and background calls

MiniMax M3

Free

1M tokens

Long-context and multimodal work

DiffusionGemma

Free

262K tokens

Fast local text generation

Every model above is available to every account. A Pro-enabled key can unlock further models such as GLM-5.2 and Kimi K2.7 Code, and bge-m3 is available for codebase indexing via /v1/embeddings. The authenticated GET /v1/models response is the source of truth for what your key can use.

See the complete Available Models list for all available models.

Support

Need help? Check out: