All tasks / Tools
Llama
Small menu-bar/tray app for macOS and Windows from the llama.cpp team that runs open models locally with chat and an OpenAI-style API.
- ggml-org and Hugging Face
- Local only
- Free, open source
Where it ranks
- AChat & assistantsApps · Local · via Gemma 4 (31B)
Tools without a benchmark of their own inherit the tier of the best model they offer.
Facts
- Vendor
- ggml-org and Hugging Face
- Approach
- Apps
- Where it runs
- Runs only on your own hardware
- Interface
- Desktop app, API, Command line
- Skill level
- Beginner: works out of the box
- Price
- Free, open source
- Pricing notes
- Free and open source.
- Source code
- Open source (MIT)
- GitHub stars
- 1.6k (18 Oct 2026)
- Last verified
- 18 Oct 2026
Price from llama.app, 18 Oct 2026.
Models it gives you
- AGemma 4 (31B)Google DeepMind · open weights
- AQwen3.8-27BAlibaba (Qwen) · open weights
- Bgpt-oss-120bOpenAI · open weights
- BQwen3.8-2.4T-A95BAlibaba (Qwen) · open weights
- Cgpt-oss-20bOpenAI · open weights