Ollama
One command to download and run a language model on your own machine. The easiest honest answer to "how do I do this without a cloud".
- Self-hosted
- Allowed
- No GPU needed
Qwen2.5 72B Instruct
A large open-weights model strong at multilingual work, widely used as a fine-tuning base.
- Open weights
- Text
- Filtered adult content
At a glance
- Made by
- Alibaba
- Family
- Qwen
- Licence
- Qwen License
- Size
- 72B parameters
- Context
- 131072 tokens
- Graphics memory
- 24 GB minimum, 48 GB recommended
- Adult content
- FilteredA filter blocks it, with varying success.
- On Hugging Face
- Qwen/Qwen2.5-72B-Instruct
Notably better than the Llama line at languages other than English, which matters if the conversation you want is not in English. As with every instruct release, the refusals are in the tune rather than the architecture.