Last updated: April 5, 2026 · Pricing & Deployment · by Daniel Ashford

What is Open Source / Open Weights?

QUICK ANSWER

LLMs whose model weights are publicly available for download and self-hosting.

Definition

Open weights LLMs are models whose trained parameters are publicly released, allowing anyone to download, run, fine-tune, and deploy them without per-token API fees.

How It Works

Major open-weight families include Llama (Meta), Qwen (Alibaba), Mistral, DeepSeek, and Gemma (Google). Licensing varies: MIT and Apache 2.0 allow unrestricted commercial use. Open models lag frontier closed models by 6-12 months but are catching up rapidly.

Example

Llama 4 405B can be downloaded and run on your own GPUs. A company processing 500M tokens/month saves approximately $15-50K/month vs commercial API pricing.

Related Terms

Self-Hosting
Running an LLM on your own hardware instead of using a cloud API.
Fine-Tuning
Customizing a pre-trained LLM on your specific data to improve performance for your use case.
Quantization
Compressing an LLM to use less memory by reducing numerical precision.
GPU (Graphics Processing Unit)
The specialized hardware that LLMs run on.

See How Models Compare

Understanding open source / open weights is important when choosing the right AI model. See how 12 models compare on our leaderboard.

View Leaderboard →Our Methodology
← Browse all 47 glossary terms
DA
Daniel Ashford
Founder & Lead Evaluator · 200+ models evaluated