What's new

Open Source AI Models

28 open-source and open-weight AI models ranked by quality. These models can be self-hosted, fine-tuned, and deployed without vendor lock-in.

28

Open Source Models

19

Language Models

10

Providers

8

Free to Use

Open Source Language Models

# Model Quality API Price
1 DeepSeek V3.2 DeepSeek Not published $0.29 / $1.14
2 Mistral Large Mistral Not published $2.00 / $6.00
3 R1 0528 DeepSeek Not published $0.50 / $2.15
4 Llama 4 Maverick Meta Not published $0.19 / $0.65
5 Llama 4 Scout Meta Not published $0.18 / $0.59
6 Mistral Small 3.1 24B Mistral Not published $0.35 / $0.56
7 Command A Cohere Not published $2.50 / $10.00
8 DeepSeek R1 DeepSeek Not published $0.70 / $2.50
9 DeepSeek V3 DeepSeek Not published $0.32 / $0.89
10 Phi 4 Microsoft Not published $0.07 / $0.14
11 Llama 3.3 70B Instruct Meta Not published $0.59 / $0.79
12 Qwen2.5 Coder 32B Instruct Alibaba Not published $0.66 / $1.00
13 Qwen2.5 72B Instruct Alibaba Not published $0.36 / $0.40
14 Command R+ (08-2024) Cohere Not published $2.50 / $10.00
15 Llama 3.1 70B Instruct Meta Not published $0.40 / $0.40
16 Llama 3.1 8B Instruct Meta Not published $0.05 / $0.08
17 Mistral Nemo Mistral Not published $0.03 / $0.03
18 Command R7B (12-2024) Cohere Not published $0.04 / $0.15
19 Command R (08-2024) Cohere Not published $0.15 / $0.60

What does "Open Source" mean for AI?

Open-source AI models publish their model weights, allowing anyone to download, run, fine-tune, and deploy them. Some models (like Meta's Llama) use custom licences with usage restrictions, while others (like Mistral and some DeepSeek models) use fully permissive licences. "Open weight" is technically more accurate for models that share weights but not training data or code, though "open source" is commonly used.