Skip to content
AI Model Radar
Apple silicon

Apple MLX models

5,841 models converted for Apple MLX, listed from Hugging Face’s mlx-community, refreshed 1h ago. Weights size and Mac memory are estimates from parameter count and quantisation.

Alert me about new Apple/MLX models

2,922 builds · page 9 of 49 · clear filters

Llama-3.2-3B-8bit

mlx-community/Llama-3.2-3B-8bit
8-bit
Text generation ↓ 249 · ♥ 3
Params
3.2B
Weights
≈ 3.4 GB
Mac memory
8 GB+

llama3.2 · updated 2024-09-26

Phi-4-mini-instruct-mlx-8Bit

from microsoft/Phi-4-mini-instruct
8-bit
Text generation · Microsoft · Phi ↓ 248 · ♥ 1
Params
3.8B
Weights
≈ 4.1 GB
Mac memory
8 GB+

mit · updated 2026-07-27

Llama-3.2-3B-bf16

mlx-community/Llama-3.2-3B-bf16
bf16
Text generation ↓ 247 · ♥ 6
Params
3.2B
Weights
≈ 6.4 GB
Mac memory
16 GB+

llama3.2 · updated 2024-09-26

HY-MT1.5-1.8B-8bit

from tencent/HY-MT1.5-1.8B
8-bit
Text generation · tencent ↓ 246 · ♥ 6
Params
1.8B
Weights
≈ 1.9 GB
Mac memory
8 GB+

updated 2026-01-03

medgemma-27b-text-it-4bit

from google/medgemma-27b-text-it
4-bit
Text generation · Google · Gemma ↓ 246 · ♥ 3
Params
28B
Weights
≈ 16.0 GB
Mac memory
24 GB+

health-ai-developer-foundations · updated 2025-06-09

Qwen3-4B-bf16

from Qwen/Qwen3-4B
bf16
Text generation · Qwen ↓ 246 · ♥ 0
Params
4.0B
Weights
≈ 8.0 GB
Mac memory
16 GB+

apache-2.0 · updated 2025-04-28

Qwen2.5-14B-Instruct-1M-8bit

from Qwen/Qwen2.5-14B-Instruct-1M
8-bit
Text generation · Qwen ↓ 245 · ♥ 12
Params
15B
Weights
≈ 15.7 GB
Mac memory
24 GB+

apache-2.0 · updated 2025-01-26

NVIDIA-Nemotron-3-Super-120B-A12B-5bit

from nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16
5-bit
Text generation · NVIDIA ↓ 243 · ♥ 3
Params
121B
Weights
≈ 83.1 GB
Mac memory
96 GB+

nvidia-nemotron-open-model-license · updated 2026-03-30

gemma-4-31B-it-qat-assistant-4bit

from google/gemma-4-31B-it-qat-q4_0-unquantized-assistant
4-bit
Text generation · Google · Gemma ↓ 243 · ♥ 0
Params
470M
Weights
≈ 264 MB
Mac memory
8 GB+

gemma · updated 2026-06-05

c4ai-command-r-v01-2bit

mlx-community/c4ai-command-r-v01-2bit
2-bit
Text generation ↓ 242 · ♥ 4
Params
—
Weights
—
Mac memory
—

updated 2024-03-12

meta-llama-Llama-4-Scout-17B-16E-8bit

from meta-llama/Llama-4-Scout-17B-16E
8-bit
Text generation · Meta · Llama ↓ 241 · ♥ 4
Params
108B
Weights
≈ 115 GB
Mac memory
192 GB+

llama4 · updated 2025-04-06

Ornith-1.5-35B-A3B-8bit

from ornith-ai/Ornith-1.5-35B-A3B
8-bit
Text generation · ornith-ai ↓ 241 · ♥ 1
Params
35B
Weights
≈ 37.7 GB
Mac memory
48 GB+

mit · updated 2026-08-21

GLM-5-8bit-MXFP8

mlx-community/GLM-5-8bit-MXFP8
MXFP8
Text generation ↓ 240 · ♥ 0
Params
744B
Weights
≈ 791 GB
Mac memory
> 512 GB

updated 2026-02-13

Hunyuan-MT-7B-4bit

from tencent/Hunyuan-MT-7B
4-bit
Text generation · tencent ↓ 240 · ♥ 0
Params
7.5B
Weights
≈ 4.2 GB
Mac memory
8 GB+

updated 2026-06-13

gemma-2-27b-it-8bit

mlx-community/gemma-2-27b-it-8bit
8-bit
Text generation ↓ 239 · ♥ 10
Params
27B
Weights
≈ 28.9 GB
Mac memory
48 GB+

gemma · updated 2024-11-06

CodeLlama-7b-Python-4bit-MLX

mlx-community/CodeLlama-7b-Python-4bit-MLX
4-bit
Text generation ↓ 238 · ♥ 14
Params
—
Weights
—
Mac memory
—

llama2 · updated 2024-02-02

Qwen2.5-Coder-32B-Instruct-abliterated-3bit

from huihui-ai/Qwen2.5-Coder-32B-Instruct-abliterated
3-bit
Text generation · huihui-ai ↓ 236 · ♥ 1
Params
33B
Weights
≈ 14.3 GB
Mac memory
24 GB+

apache-2.0 · updated 2025-01-06

c4ai-command-r-v01-4bit

mlx-community/c4ai-command-r-v01-4bit
4-bit
Text generation ↓ 233 · ♥ 23
Params
35B
Weights
≈ 22.7 GB
Mac memory
32 GB+

updated 2024-03-14

Hy4-preview-4bit

from tencent/Hy4-preview
4-bit
Text generation · tencent ↓ 233 · ♥ 2
Params
770B
Weights
≈ 433 GB
Mac memory
512 GB+

apache-2.0 · updated 2026-08-28

LFM2.5-230M-OptiQ-4bit

from LiquidAI/LFM2.5-230M
4-bit
Text generation · LiquidAI ↓ 233 · ♥ 0
Params
230M
Weights
≈ 129 MB
Mac memory
8 GB+

lfm1.0 · updated 2026-10-14

Phi-4-mini-instruct-6bit

from microsoft/Phi-4-mini-instruct
6-bit
Text generation · Microsoft · Phi ↓ 232 · ♥ 3
Params
3.8B
Weights
≈ 3.1 GB
Mac memory
8 GB+

mit · updated 2025-03-05

Qwen2.5-14B-Instruct-1M-4bit

from Qwen/Qwen2.5-14B-Instruct-1M
4-bit
Text generation · Qwen ↓ 231 · ♥ 2
Params
15B
Weights
≈ 8.3 GB
Mac memory
16 GB+

apache-2.0 · updated 2025-01-26

Devstral-Small-2505-6bit

from mistralai/Devstral-Small-2505
6-bit
Text generation · Mistral ↓ 230 · ♥ 1
Params
24B
Weights
≈ 19.2 GB
Mac memory
24 GB+

apache-2.0 · updated 2025-05-21

GLM-5.1-MXFP4-Q8

mlx-community/GLM-5.1-MXFP4-Q8
MXFP4
Text generation ↓ 229 · ♥ 4
Params
744B
Weights
≈ 419 GB
Mac memory
512 GB+

updated 2026-04-08

Meta-Llama-3.1-8B-Instruct-abliterated-4bit

from meta-llama/Llama-3.1-8B-Instruct
4-bit
Text generation · Meta · Llama ↓ 229 · ♥ 0
Params
8.0B
Weights
≈ 4.5 GB
Mac memory
8 GB+

llama3.1 · updated 2024-08-03

Qwen2-57B-A14B-Instruct-4bit

mlx-community/Qwen2-57B-A14B-Instruct-4bit
4-bit
Text generation ↓ 228 · ♥ 2
Params
57B
Weights
≈ 32.3 GB
Mac memory
48 GB+

apache-2.0 · updated 2024-06-06

Ministral-3-8B-Instruct-2512

from mistralai/Ministral-3-8B-Instruct-2512
Text generation · Mistral ↓ 227 · ♥ 2
Params
8.5B
Weights
≈ 17.0 GB
Mac memory
24 GB+

apache-2.0 · updated 2025-12-03

Qwen2.5-Coder-1.5B-8bit

from Qwen/Qwen2.5-1.5B
8-bit
Text generation · Qwen ↓ 227 · ♥ 0
Params
1.5B
Weights
≈ 1.6 GB
Mac memory
8 GB+

apache-2.0 · updated 2024-09-18

granite-4.1-3b-8bit

from ibm-granite/granite-4.1-3b
8-bit
Text generation · IBM · Granite ↓ 227 · ♥ 0
Params
3.4B
Weights
≈ 3.6 GB
Mac memory
8 GB+

apache-2.0 · updated 2026-05-01

Qwen3-4B-Instruct-2507-gabliterated-4bit

from Goekdeniz-Guelmez/Qwen3-4B-Instruct-2507-gabliterated
4-bit
Text generation · Goekdeniz-Guelmez ↓ 225 · ♥ 0
Params
4.0B
Weights
≈ 2.3 GB
Mac memory
8 GB+

updated 2026-01-06

gemma-3-text-12b-it-4bit

from mlx-community/gemma-3-12b-it-bf16
4-bit
Text generation ↓ 225 · ♥ 0
Params
13B
Weights
≈ 7.2 GB
Mac memory
16 GB+

gemma · updated 2025-03-18

GLM-5-4bit

from zai-org/GLM-5
4-bit
Text generation · zai-org ↓ 224 · ♥ 8
Params
744B
Weights
≈ 419 GB
Mac memory
512 GB+

mit · updated 2026-02-12

Devstral-Small-2505-8bit

from mistralai/Devstral-Small-2505
8-bit
Text generation · Mistral ↓ 224 · ♥ 2
Params
24B
Weights
≈ 25.0 GB
Mac memory
32 GB+

apache-2.0 · updated 2025-05-21

opt-125m-chat-mlx-4bit

from facebook/opt-125m
4-bit
Text generation · facebook ↓ 224 · ♥ 0
Params
164M
Weights
≈ 92 MB
Mac memory
8 GB+

other · updated 2026-10-14

DeepSeek-R1-Distill-Llama-70B-3bit

from deepseek-ai/DeepSeek-R1-Distill-Llama-70B
3-bit
Text generation · DeepSeek ↓ 223 · ♥ 5
Params
71B
Weights
≈ 30.9 GB
Mac memory
48 GB+

updated 2025-02-26

Falcon3-10B-Instruct-4bit

from tiiuae/Falcon3-10B-Instruct
4-bit
Text generation · TII · Falcon ↓ 222 · ♥ 0
Params
10B
Weights
≈ 5.8 GB
Mac memory
16 GB+

falcon-llm-license · updated 2024-12-17

Josiefied-LFM2.5-1.2B-Instruct-abliterated-4bit

from Goekdeniz-Guelmez/Josiefied-LFM2.5-1.2B-Instruct-abliterated
4-bit
Text generation · Goekdeniz-Guelmez ↓ 222 · ♥ 0
Params
1.2B
Weights
≈ 659 MB
Mac memory
8 GB+

updated 2026-01-26

Qwen3-4B-Instruct-2507-6bit

from Qwen/Qwen3-4B-Instruct-2507
6-bit
Text generation · Qwen ↓ 221 · ♥ 1
Params
4.0B
Weights
≈ 3.3 GB
Mac memory
8 GB+

apache-2.0 · updated 2025-08-06

LFM2.5-2.6B-8bit

from LiquidAI/LFM2.5-2.6B
8-bit
Text generation · LiquidAI ↓ 220 · ♥ 3
Params
2.7B
Weights
≈ 2.9 GB
Mac memory
8 GB+

lfm1.0 · updated 2026-08-13

Phi-4-mini-instruct-mlx-fp16

from microsoft/Phi-4-mini-instruct
fp16
Text generation · Microsoft · Phi ↓ 220 · ♥ 1
Params
3.8B
Weights
≈ 7.7 GB
Mac memory
16 GB+

mit · updated 2026-07-27

DeepSeek-R1-Distill-Llama-70B-6bit

from deepseek-ai/DeepSeek-R1-Distill-Llama-70B
6-bit
Text generation · DeepSeek ↓ 219 · ♥ 4
Params
71B
Weights
≈ 57.3 GB
Mac memory
96 GB+

updated 2025-02-26

moonshotai_Kimi-K2-Instruct-mlx-3bit

from moonshotai/Kimi-K2-Instruct
3-bit
Text generation · Moonshot · Kimi ↓ 219 · ♥ 2
Params
1.0T
Weights
≈ 449 GB
Mac memory
512 GB+

modified-mit · updated 2025-08-23

Qwen3.5-27B-4bit-DWQ

mlx-community/Qwen3.5-27B-4bit-DWQ
4-bit DWQ
Text generation ↓ 218 · ♥ 6
Params
27B
Weights
≈ 15.1 GB
Mac memory
24 GB+

updated 2026-03-24

Phi-4-reasoning-plus-4bit

from microsoft/Phi-4-reasoning-plus
4-bit
Text generation · Microsoft · Phi ↓ 218 · ♥ 1
Params
15B
Weights
≈ 8.2 GB
Mac memory
16 GB+

mit · updated 2025-05-01

Unsloth-DeepSeek-R1-Distill-Qwen-32B-4bit

from unsloth/DeepSeek-R1-Distill-Qwen-32B
4-bit
Text generation · unsloth ↓ 217 · ♥ 2
Params
33B
Weights
≈ 18.4 GB
Mac memory
24 GB+

apache-2.0 · updated 2025-02-26

granite-4.0-h-350m-4bit

from ibm-granite/granite-4.0-h-350m
4-bit
Text generation · IBM · Granite ↓ 217 · ♥ 0
Params
340M
Weights
≈ 192 MB
Mac memory
8 GB+

apache-2.0 · updated 2025-10-28

Qwen3-30B-A3B-Instruct-2507-8bit

from Qwen/Qwen3-30B-A3B-Instruct-2507
8-bit
Text generation · Qwen ↓ 216 · ♥ 2
Params
31B
Weights
≈ 32.4 GB
Mac memory
48 GB+

apache-2.0 · updated 2025-07-29

glm-4.7-flash-abliterated-8bit

from huihui-ai/Huihui-GLM-4.7-Flash-abliterated
8-bit
Text generation · huihui-ai ↓ 216 · ♥ 0
Params
30B
Weights
≈ 31.8 GB
Mac memory
48 GB+

updated 2026-03-30

CodeLlama-70b-Instruct-hf-4bit-MLX

mlx-community/CodeLlama-70b-Instruct-hf-4bit-MLX
4-bit
Text generation ↓ 214 · ♥ 25
Params
—
Weights
—
Mac memory
—

llama2 · updated 2024-01-30

OpenCodeInterpreter-DS-33B-hf-4bit-mlx

mlx-community/OpenCodeInterpreter-DS-33B-hf-4bit-mlx
4-bit
Text generation ↓ 214 · ♥ 7
Params
33B
Weights
≈ 19.1 GB
Mac memory
24 GB+

updated 2024-02-24

Qwen3-8B-4bit-DWQ-053125

from Qwen/Qwen3-8B
4-bit DWQ
Text generation · Qwen ↓ 214 · ♥ 3
Params
8.2B
Weights
≈ 4.6 GB
Mac memory
8 GB+

apache-2.0 · updated 2025-06-01

DeepSeek-R1-Distill-Qwen-32B-Japanese-8bit

from cyberagent/DeepSeek-R1-Distill-Qwen-32B-Japanese
8-bit
Text generation · cyberagent ↓ 214 · ♥ 2
Params
33B
Weights
≈ 34.8 GB
Mac memory
48 GB+

mit · updated 2025-01-27

DeepSeek-R1-Distill-Qwen-32B-abliterated

from huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated
Text generation · huihui-ai ↓ 213 · ♥ 3
Params
33B
Weights
≈ 65.5 GB
Mac memory
96 GB+

updated 2025-02-20

Phi-3.5-MoE-instruct-8bit

mlx-community/Phi-3.5-MoE-instruct-8bit
8-bit
Text generation ↓ 213 · ♥ 3
Params
42B
Weights
≈ 44.5 GB
Mac memory
64 GB+

mit · updated 2024-08-24

Qwen3-14B-4bit-DWQ-053125

from Qwen/Qwen3-14B
4-bit DWQ
Text generation · Qwen ↓ 212 · ♥ 8
Params
15B
Weights
≈ 8.3 GB
Mac memory
16 GB+

apache-2.0 · updated 2025-06-02

BTL-4-OptiQ-4bit

from badtheorylabs/BTL-4
4-bit
Text generation · badtheorylabs ↓ 211 · ♥ 0
Params
35B
Weights
≈ 19.5 GB
Mac memory
24 GB+

apache-2.0 · updated 2026-09-15

NVIDIA-Nemotron-3-Nano-30B-A3B-OptiQ-4bit

from nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
4-bit
Text generation · NVIDIA ↓ 210 · ♥ 2
Params
6.0B
Weights
≈ 4.8 GB
Mac memory
8 GB+

nvidia-open-model-license · updated 2026-09-14