4-bit Open on Hugging Face
Parameters
40B
Weights (est.)
≈ 22.4 GB
Mac memory (est.)
32 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/flint-oQ4e --prompt "Hello"
Details
- Base model
- RecursiveSI/flint (quantized)
- Vendor
- RecursiveSI
- Task
- Text generation
- Quantisation
- 4-bit
- Licence
- apache-2.0
- Downloads
- 1.2K in the last 30 days · 1.2K all time · 2 likes
- Published
- 2026-10-27 · updated 2026-10-27 (15d ago)
- Tags
- flint, mlx-vlm, omlx, quantized, code, agent
Other MLX builds of RecursiveSI/flint
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| flint-oQ2e | 2-bit | ≈ 12.5 GB | 16 GB+ | 406 |
| flint-oQ2.7e | 2-bit | ≈ 12.5 GB | 16 GB+ | 392 |
| flint-oQ2.5e | 2-bit | ≈ 12.5 GB | 16 GB+ | 373 |
| flint-oQ3e | 3-bit | ≈ 17.4 GB | 24 GB+ | 428 |
| flint-OptiQ-4bit | 4-bit | ≈ 22.4 GB | 32 GB+ | 1.9K |
| flint-oQ4e (this one) | 4-bit | ≈ 22.4 GB | 32 GB+ | 1.2K |
| flint-oQ5e | 5-bit | ≈ 27.4 GB | 48 GB+ | 436 |
| flint-oQ6e | 6-bit | ≈ 32.4 GB | 48 GB+ | 551 |
| flint-oQ8e | 8-bit | ≈ 42.3 GB | 64 GB+ | 804 |