4-bit Open on Hugging Face
Parameters
3.2B
Weights (est.)
≈ 1.8 GB
Mac memory (est.)
8 GB+
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Run it
With mlx-lm on an Apple silicon Mac:
pip install mlx-lmmlx_lm.generate --model mlx-community/kanana-2-3b-instruct-4bit --prompt "Hello"
Details
- Base model
- kakaocorp/kanana-2-3b-instruct (quantized)
- Vendor
- kakaocorp
- Task
- Text generation
- Quantisation
- 4-bit
- Licence
- kanana-open-license
- Downloads
- 163 in the last 30 days · 496 all time · 0 likes
- Published
- 2026-08-04 · updated 2026-08-06 (3mo ago)
- Languages
- ko, en
- Tags
- qwen3, kanana, korean
Other MLX builds of kakaocorp/kanana-2-3b-instruct
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| kanana-2-3b-instruct-4bit (this one) | 4-bit | ≈ 1.8 GB | 8 GB+ | 163 |
| kanana-2-3b-instruct-mixed_4_6 | 4-bit | ≈ 1.8 GB | 8 GB+ | 35 |
| kanana-2-3b-instruct-6bit | 6-bit | ≈ 2.6 GB | 8 GB+ | 36 |
| kanana-2-3b-instruct-8bit | 8-bit | ≈ 3.4 GB | 8 GB+ | 70 |