-
mlx-community/gemma-3n-E4B-it-bf16
Image-Text-to-Text • 8B • Updated • 356 • 1 -
mlx-community/gemma-3n-E2B-it-bf16
Image-Text-to-Text • 6B • Updated • 260 -
mlx-community/gemma-3n-E4B-bf16
Image-Text-to-Text • 8B • Updated • 271 • 4 -
mlx-community/gemma-3n-E2B-bf16
Image-Text-to-Text • 6B • Updated • 350 • 2
AI & ML interests
None defined yet.
Recent Activity
View all activity
This collection houses BitNet-1.58, Falcon3-1.58 and Falcon-E quants.
-
mlx-community/bitnet-b1.58-2B-4T
Text Generation • 0.8B • Updated • 270 -
mlx-community/bitnet-b1.58-2B-4T-4bit
Text Generation • 0.6B • Updated • 266 -
mlx-community/bitnet-b1.58-2B-4T-8bit
Text Generation • 0.6B • Updated • 148 -
mlx-community/bitnet-b1.58-2B-4T-6bit
Text Generation • 0.6B • Updated • 77
High-quality 4-bit quants of the Qwen3 model family.
-
mlx-community/Qwen3-14B-4bit-DWQ-053125
Text Generation • 2B • Updated • 491 • 2 -
mlx-community/Qwen3-8B-4bit-DWQ-053125
Text Generation • 1B • Updated • 389 • 1 -
mlx-community/Qwen3-4B-4bit-DWQ-053125
Text Generation • 0.6B • Updated • 2.8k • 1 -
mlx-community/Qwen3-1.7B-4bit-DWQ-053125
Text Generation • 0.3B • Updated • 2.14k
-
mlx-community/AceReason-Nemotron-7B-4bit
Text Generation • 1B • Updated • 96 -
mlx-community/AceReason-Nemotron-7B-8bit
Text Generation • 2B • Updated • 38 -
mlx-community/AceReason-Nemotron-7B-bf16
Text Generation • 8B • Updated • 34 -
mlx-community/AceReason-Nemotron-14B-4bit
Text Generation • 2B • Updated • 145
Collection of Gemma 3 variants for performance on medical text and image comprehension to accelerate building healthcare-based AI applications.
-
mlx-community/medgemma-4b-it-4bit
Image-Text-to-Text • 0.9B • Updated • 306 • 2 -
mlx-community/medgemma-4b-it-6bit
Image-Text-to-Text • 1B • Updated • 162 • 1 -
mlx-community/medgemma-4b-it-8bit
Image-Text-to-Text • 1B • Updated • 233 • 1 -
mlx-community/medgemma-4b-it-bf16
Image-Text-to-Text • 5B • Updated • 1.88k • 1
Nvidia's ASR models, now in MLX!
-
mlx-community/parakeet-ctc-0.6b
Automatic Speech Recognition • 0.6B • Updated • 97 • 1 -
mlx-community/parakeet-rnnt-0.6b
Automatic Speech Recognition • 0.6B • Updated • 63 -
mlx-community/parakeet-ctc-1.1b
Automatic Speech Recognition • 1B • Updated • 83 • 1 -
mlx-community/parakeet-rnnt-1.1b
Automatic Speech Recognition • 1B • Updated • 98 • 1
The GLM-4 and Z1 series are powerful open-source language models excelling in reasoning, code, and complex tasks.
-
mlx-community/GLM-Z1-32B-0414-4bit
Text Generation • 5B • Updated • 1.82k • 2 -
mlx-community/GLM-4-32B-0414-4bit
Text Generation • 5B • Updated • 2.02k • 4 -
mlx-community/GLM-4-32B-Base-0414-8bit
Text Generation • 9B • Updated • 110 -
mlx-community/GLM-4-32B-Base-0414-6bit
Text Generation • 7B • Updated • 73
-
mlx-community/Llama-4-Scout-17B-16E-Instruct-4bit
Image-Text-to-Text • Updated • 3.56k • 8 -
mlx-community/Llama-4-Scout-17B-16E-Instruct-6bit
Image-Text-to-Text • Updated • 2.49k • 5 -
mlx-community/Llama-4-Scout-17B-16E-Instruct-8bit
Image-Text-to-Text • Updated • 2.6k • 1 -
mlx-community/Llama-4-Maverick-17B-16E-Instruct-4bit
Text Generation • 63B • Updated • 2.79k • 7
A collection of lightweight, state-of-the-art open models built from the same research and technology that powers the Gemini 2.0 models
-
mlx-community/gemma-3-4b-it-8bit
Image-Text-to-Text • 2B • Updated • 9.94k • 3 -
mlx-community/gemma-3-4b-pt-4bit
Image-Text-to-Text • 1B • Updated • 1.84k • 3 -
mlx-community/gemma-3-4b-it-bf16
Image-Text-to-Text • 5B • Updated • 2.4k -
mlx-community/gemma-3-4b-pt-6bit
Image-Text-to-Text • 1B • Updated • 24
-
mlx-community/OLMoE-1B-7B-0125-Instruct
Text Generation • 7B • Updated • 16 -
mlx-community/OLMoE-1B-7B-0125-Instruct-8bit
Text Generation • 2B • Updated • 11 -
mlx-community/OLMoE-1B-7B-0125-Instruct-6bit
Text Generation • 2B • Updated • 11 -
mlx-community/OLMoE-1B-7B-0125-Instruct-4bit
Text Generation • 1B • Updated • 16 • 1
-
mlx-community/olmOCR-7B-0225-preview-bf16
Image-Text-to-Text • 8B • Updated • 106 • 3 -
mlx-community/olmOCR-7B-0225-preview-4bit
Image-Text-to-Text • 2B • Updated • 1.86k • 1 -
mlx-community/olmOCR-7B-0225-preview-6bit
Image-Text-to-Text • 2B • Updated • 38 -
mlx-community/olmOCR-7B-0225-preview-8bit
Updated
FuseAI is attempting to merge CoT models to achieve newer models that are more than the sum of their parts.
-
mlx-community/Qwen2.5-VL-72B-Instruct-8bit
Image-Text-to-Text • 21B • Updated • 139 • 2 -
mlx-community/Qwen2.5-VL-72B-Instruct-6bit
Image-Text-to-Text • 16B • Updated • 58 • 1 -
mlx-community/Qwen2.5-VL-72B-Instruct-4bit
Image-Text-to-Text • 12B • Updated • 2.03k • 6 -
mlx-community/Qwen2.5-VL-72B-Instruct-3bit
Image-Text-to-Text • 10B • Updated • 144 • 4
Kyutai's Helium-1 2B Model, outperforming other state of the art small models.
-
mlx-community/helium-1-preview-2b-float32
Text Generation • 2B • Updated • 9 -
mlx-community/helium-1-preview-2b
Text Generation • 2B • Updated • 58 -
mlx-community/helium-1-preview-2b-8bit
Text Generation • 0.6B • Updated • 25 • 1 -
mlx-community/helium-1-preview-2b-4bit
Text Generation • 0.3B • Updated • 12 • 1
-
mlx-community/deepseek-vl2-6bit
Image-Text-to-Text • 6B • Updated • 31 • 1 -
mlx-community/deepseek-vl2-small-4bit
Image-Text-to-Text • 3B • Updated • 81 -
mlx-community/deepseek-vl2-4bit
Image-Text-to-Text • 4B • Updated • 78 • 1 -
mlx-community/deepseek-vl2-small-6bit
Image-Text-to-Text • 4B • Updated • 32
-
mlx-community/Llama-3.3-70B-Instruct-8bit
Text Generation • 20B • Updated • 2.55k • 12 -
mlx-community/Llama-3.3-70B-Instruct-6bit
Text Generation • 15B • Updated • 215 • 5 -
mlx-community/Llama-3.3-70B-Instruct-3bit
Text Generation • 9B • Updated • 321 • 6 -
mlx-community/Llama-3.3-70B-Instruct-4bit
Text Generation • 11B • Updated • 5.03k • 28
Falcon Mamba models compatible with MLX
Google’s Code-Gemma
The Qwen 2.5 models are a series of AI models trained on 18 trillion tokens, supporting 29 languages and offering advanced features such as instructio
-
mlx-community/Qwen2.5-72B-Instruct-bf16
Text Generation • 73B • Updated • 25 -
mlx-community/Qwen2.5-72B-Instruct-8bit
Text Generation • 20B • Updated • 52 • 4 -
mlx-community/Qwen2.5-72B-Instruct-4bit
Text Generation • 11B • Updated • 1.89k • 6 -
mlx-community/Qwen2.5-32B-Instruct-bf16
Text Generation • 33B • Updated • 27
OpenAI Whisper speech recognition models in MLX format
A series of smol LLMs: 135M, 360M and 1.7B.
-
mlx-community/Meta-Llama-3.1-70B-bf16
Text Generation • 71B • Updated • 52 • 3 -
mlx-community/Meta-Llama-3.1-70B-Instruct-bf16
Text Generation • 71B • Updated • 35 • 2 -
mlx-community/Meta-Llama-3.1-8B-Instruct-bf16
Text Generation • 8B • Updated • 119 • • 3 -
mlx-community/Meta-Llama-3.1-8B-Instruct-8bit
Text Generation • 2B • Updated • 2.18k • 9
Meta goes small with Llama3.2, both text only 1B and 3B, and the 11B Vision models.
-
mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated
Image-Text-to-Text • 11B • Updated • 2.25k • 7 -
mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated-8-bit
Image-Text-to-Text • 3B • Updated • 151 -
mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated-4-bit
Image-Text-to-Text • 2B • Updated • 186 • 1 -
mlx-community/Llama-3.2-11B-Vision-Instruct-8bit
Image-Text-to-Text • 3B • Updated • 2.58k • 10
This collection houses Nanonets-OCR-s
-
mlx-community/DeepSeek-R1-0528-4bit
Text Generation • 105B • Updated • 5.02k • 13 -
mlx-community/DeepSeek-R1-0528-Qwen3-8B-4bit
Text Generation • 1B • Updated • 3.96k • 4 -
mlx-community/DeepSeek-R1-0528-Qwen3-8B-4bit-DWQ
Text Generation • 1B • Updated • 4.03k • 8 -
mlx-community/DeepSeek-R1-0528-Qwen3-8B-8bit
Text Generation • 2B • Updated • 2.84k • 1
-
mlx-community/Devstral-Small-2505-3bit
Text Generation • 3B • Updated • 474 • 1 -
mlx-community/Devstral-Small-2505-4bit
Text Generation • 4B • Updated • 445 -
mlx-community/Devstral-Small-2505-6bit
Text Generation • Updated • 393 -
mlx-community/Devstral-Small-2505-8bit
Text Generation • Updated • 924 • 1
-
mlx-community/Llama-OuteTTS-1.0-1B-fp16
Text-to-Speech • 1B • Updated • 185 • 2 -
mlx-community/Llama-OuteTTS-1.0-1B-4bit
Text-to-Speech • 0.2B • Updated • 118 • 1 -
mlx-community/Llama-OuteTTS-1.0-1B-8bit
Text-to-Speech • 0.4B • Updated • 37 • 1 -
mlx-community/Llama-OuteTTS-1.0-1B-6bit
Text-to-Speech • 0.3B • Updated • 17
Gemma 3 distilled weight quantized (DWQ) models
-
mlx-community/gemma-3-4b-it-4bit-DWQ
Text Generation • 0.7B • Updated • 2.02k -
mlx-community/gemma-3-12b-it-4bit-DWQ
Text Generation • 2B • Updated • 2.21k • 2 -
mlx-community/gemma-3-1b-it-4bit-DWQ
Text Generation • 0.2B • Updated • 1.83k -
mlx-community/gemma-3-27b-it-4bit-DWQ
Text Generation • 4B • Updated • 2.32k • 2
Abliterated, and further fine-tuned to be the most uncensored models available. Now in MLX
-
mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-bf16
Text Generation • 31B • Updated • 91 -
mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-8bit
Text Generation • 31B • Updated • 137 -
mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-6bit
Text Generation • 31B • Updated • 111 -
mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-4bit
Text Generation • 31B • Updated • 194
Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory.
-
mlx-community/gemma-3-27b-it-qat-bf16
Image-Text-to-Text • Updated • 2.22k • 4 -
mlx-community/gemma-3-27b-it-qat-8bit
Image-Text-to-Text • Updated • 2.29k • 5 -
mlx-community/gemma-3-27b-it-qat-6bit
Image-Text-to-Text • Updated • 164 -
mlx-community/gemma-3-27b-it-qat-4bit
Image-Text-to-Text • Updated • 182k • 18
-
mlx-community/answerdotai-ModernBERT-base-8bit
Fill-Mask • 0.1B • Updated • 28 -
mlx-community/answerdotai-ModernBERT-base-4bit
Fill-Mask • 0.0B • Updated • 29 -
mlx-community/answerdotai-ModernBERT-base-bf16
Fill-Mask • 0.2B • Updated • 20 -
mlx-community/answerdotai-ModernBERT-Large-Instruct-4bit
Fill-Mask • 0.1B • Updated • 14
Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers amazing quality.
-
mlx-community/Qwen2.5-7B-Instruct-1M-4bit
Text Generation • 1B • Updated • 2.34k • 9 -
mlx-community/Qwen2.5-7B-Instruct-1M-6bit
Text Generation • 2B • Updated • 38 • 2 -
mlx-community/Qwen2.5-7B-Instruct-1M-3bit
Text Generation • 1.0B • Updated • 13 -
mlx-community/Qwen2.5-7B-Instruct-1M-8bit
Text Generation • 2B • Updated • 60 • 3
Convert HTML content to LLM-friendly Markdown/JSON content
-
mlx-community/QVQ-72B-Preview-4bit
Image-Text-to-Text • 11B • Updated • 48 • 7 -
mlx-community/QVQ-72B-Preview-6bit
Image-Text-to-Text • 16B • Updated • 14 • 2 -
mlx-community/QVQ-72B-Preview-3bit
Image-Text-to-Text • 9B • Updated • 30 • 5 -
mlx-community/QVQ-72B-Preview-8bit
Image-Text-to-Text • 21B • Updated • 20 • 3
The best uncensored models
-
mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1
Text Generation • 8B • Updated • 24 -
mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-8bit
Text Generation • 2B • Updated • 40 -
mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-6bit
Text Generation • 2B • Updated • 25 -
mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-4bit
Text Generation • 1B • Updated • 71 • 1
EXAONE 3.5, a collection of instruction-tuned bilingual generative models ranging from 2.4B to 32B parameters, developed by LG AI.
-
mlx-community/paligemma2-3b-ft-docci-448-8bit
Image-Text-to-Text • 0.9B • Updated • 16 -
mlx-community/paligemma2-3b-ft-docci-448-6bit
Image-Text-to-Text • 0.7B • Updated • 41 -
mlx-community/paligemma2-3b-ft-docci-448-bf16
Image-Text-to-Text • 3B • Updated • 18 • 1 -
mlx-community/paligemma2-10b-ft-docci-448-bf16
Image-Text-to-Text • 10B • Updated • 46 • 3
-
mlx-community/SmolVLM-Instruct-4bit
Image-Text-to-Text • 0.5B • Updated • 1.91k • 4 -
mlx-community/SmolVLM-Instruct-6bit
Image-Text-to-Text • 0.6B • Updated • 15 -
mlx-community/SmolVLM-Instruct-8bit
Image-Text-to-Text • 0.7B • Updated • 75 • 9 -
mlx-community/SmolVLM-Instruct-bf16
Image-Text-to-Text • 2B • Updated • 40 • 4
-
mlx-community/Florence-2-base-ft-4bit
Image-Text-to-Text • 0.0B • Updated • 100 • 2 -
mlx-community/Florence-2-large-ft-bf16
Image-Text-to-Text • 0.8B • Updated • 70 • 1 -
mlx-community/Florence-2-base-ft-bf16
Image-Text-to-Text • 0.3B • Updated • 24 • 1 -
mlx-community/Florence-2-base-ft-8bit
Image-Text-to-Text • 0.1B • Updated • 29 • 1
Code-specific model series based on Qwen2.5
-
mlx-community/Qwen2.5-Coder-32B-Instruct-8bit
Text Generation • 9B • Updated • 265 • 10 -
mlx-community/Qwen2.5-Coder-14B-Instruct-4bit
Text Generation • 2B • Updated • 2.3k • 4 -
mlx-community/Qwen2.5-Coder-14B-Instruct-bf16
Text Generation • 15B • Updated • 41 • 2 -
mlx-community/Qwen2.5-Coder-3B-Instruct-8bit
Text Generation • 0.9B • Updated • 173
A collection of Neversleep's RP focused Lumimaid LLMs.
Qwen1.5 is the improved version of Qwen, the large language model series developed by Alibaba Cloud.
-
mlx-community/Qwen1.5-1.8B-Chat-4bit
Text Generation • 0.5B • Updated • 50 • 2 -
mlx-community/Qwen1.5-0.5B-Chat-4bit
Text Generation • 0.1B • Updated • 4.21k • 4 -
mlx-community/Qwen1.5-14B-Chat-4bit
Text Generation • 3B • Updated • 23 • 1 -
mlx-community/Qwen1.5-7B-Chat-4bit
Text Generation • 2B • Updated • 118 • 2
-
mlx-community/Meta-Llama-3-8B-Instruct-4bit
Text Generation • 2B • Updated • 10.3k • 78 -
mlx-community/Meta-Llama-3-8B-4bit
Text Generation • 2B • Updated • 113 • 8 -
mlx-community/Meta-Llama-Guard-2-8B-4bit
Text Generation • 2B • Updated • 27 -
mlx-community/Meta-Llama-3-70B-4bit
Text Generation • 11B • Updated • 131 • 9
-
mlx-community/Phi-3-mini-4k-instruct-4bit
Text Generation • 0.6B • Updated • 2.08k • 11 -
mlx-community/Phi-3-mini-128k-instruct-4bit
Text Generation • 0.6B • Updated • 1.91k • 12 -
mlx-community/Phi-3-mini-128k-instruct-8bit
Text Generation • 1B • Updated • 75 • 10 -
mlx-community/Phi-3-mini-4k-instruct-8bit
Text Generation • 1B • Updated • 89 • 2
A family of Open-source Efficient Language Models from Apple.
Mamba is a new LLM architecture that integrates the Structured State Space sequence model to manage lengthy data sequences.
EnCodec models in MLX
-
mlx-community/gemma-3n-E4B-it-bf16
Image-Text-to-Text • 8B • Updated • 356 • 1 -
mlx-community/gemma-3n-E2B-it-bf16
Image-Text-to-Text • 6B • Updated • 260 -
mlx-community/gemma-3n-E4B-bf16
Image-Text-to-Text • 8B • Updated • 271 • 4 -
mlx-community/gemma-3n-E2B-bf16
Image-Text-to-Text • 6B • Updated • 350 • 2
This collection houses Nanonets-OCR-s
This collection houses BitNet-1.58, Falcon3-1.58 and Falcon-E quants.
-
mlx-community/bitnet-b1.58-2B-4T
Text Generation • 0.8B • Updated • 270 -
mlx-community/bitnet-b1.58-2B-4T-4bit
Text Generation • 0.6B • Updated • 266 -
mlx-community/bitnet-b1.58-2B-4T-8bit
Text Generation • 0.6B • Updated • 148 -
mlx-community/bitnet-b1.58-2B-4T-6bit
Text Generation • 0.6B • Updated • 77
High-quality 4-bit quants of the Qwen3 model family.
-
mlx-community/Qwen3-14B-4bit-DWQ-053125
Text Generation • 2B • Updated • 491 • 2 -
mlx-community/Qwen3-8B-4bit-DWQ-053125
Text Generation • 1B • Updated • 389 • 1 -
mlx-community/Qwen3-4B-4bit-DWQ-053125
Text Generation • 0.6B • Updated • 2.8k • 1 -
mlx-community/Qwen3-1.7B-4bit-DWQ-053125
Text Generation • 0.3B • Updated • 2.14k
-
mlx-community/DeepSeek-R1-0528-4bit
Text Generation • 105B • Updated • 5.02k • 13 -
mlx-community/DeepSeek-R1-0528-Qwen3-8B-4bit
Text Generation • 1B • Updated • 3.96k • 4 -
mlx-community/DeepSeek-R1-0528-Qwen3-8B-4bit-DWQ
Text Generation • 1B • Updated • 4.03k • 8 -
mlx-community/DeepSeek-R1-0528-Qwen3-8B-8bit
Text Generation • 2B • Updated • 2.84k • 1
-
mlx-community/AceReason-Nemotron-7B-4bit
Text Generation • 1B • Updated • 96 -
mlx-community/AceReason-Nemotron-7B-8bit
Text Generation • 2B • Updated • 38 -
mlx-community/AceReason-Nemotron-7B-bf16
Text Generation • 8B • Updated • 34 -
mlx-community/AceReason-Nemotron-14B-4bit
Text Generation • 2B • Updated • 145
-
mlx-community/Devstral-Small-2505-3bit
Text Generation • 3B • Updated • 474 • 1 -
mlx-community/Devstral-Small-2505-4bit
Text Generation • 4B • Updated • 445 -
mlx-community/Devstral-Small-2505-6bit
Text Generation • Updated • 393 -
mlx-community/Devstral-Small-2505-8bit
Text Generation • Updated • 924 • 1
Collection of Gemma 3 variants for performance on medical text and image comprehension to accelerate building healthcare-based AI applications.
-
mlx-community/medgemma-4b-it-4bit
Image-Text-to-Text • 0.9B • Updated • 306 • 2 -
mlx-community/medgemma-4b-it-6bit
Image-Text-to-Text • 1B • Updated • 162 • 1 -
mlx-community/medgemma-4b-it-8bit
Image-Text-to-Text • 1B • Updated • 233 • 1 -
mlx-community/medgemma-4b-it-bf16
Image-Text-to-Text • 5B • Updated • 1.88k • 1
-
mlx-community/Llama-OuteTTS-1.0-1B-fp16
Text-to-Speech • 1B • Updated • 185 • 2 -
mlx-community/Llama-OuteTTS-1.0-1B-4bit
Text-to-Speech • 0.2B • Updated • 118 • 1 -
mlx-community/Llama-OuteTTS-1.0-1B-8bit
Text-to-Speech • 0.4B • Updated • 37 • 1 -
mlx-community/Llama-OuteTTS-1.0-1B-6bit
Text-to-Speech • 0.3B • Updated • 17
Gemma 3 distilled weight quantized (DWQ) models
-
mlx-community/gemma-3-4b-it-4bit-DWQ
Text Generation • 0.7B • Updated • 2.02k -
mlx-community/gemma-3-12b-it-4bit-DWQ
Text Generation • 2B • Updated • 2.21k • 2 -
mlx-community/gemma-3-1b-it-4bit-DWQ
Text Generation • 0.2B • Updated • 1.83k -
mlx-community/gemma-3-27b-it-4bit-DWQ
Text Generation • 4B • Updated • 2.32k • 2
Nvidia's ASR models, now in MLX!
-
mlx-community/parakeet-ctc-0.6b
Automatic Speech Recognition • 0.6B • Updated • 97 • 1 -
mlx-community/parakeet-rnnt-0.6b
Automatic Speech Recognition • 0.6B • Updated • 63 -
mlx-community/parakeet-ctc-1.1b
Automatic Speech Recognition • 1B • Updated • 83 • 1 -
mlx-community/parakeet-rnnt-1.1b
Automatic Speech Recognition • 1B • Updated • 98 • 1
Abliterated, and further fine-tuned to be the most uncensored models available. Now in MLX
-
mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-bf16
Text Generation • 31B • Updated • 91 -
mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-8bit
Text Generation • 31B • Updated • 137 -
mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-6bit
Text Generation • 31B • Updated • 111 -
mlx-community/Josiefied-Qwen3-30B-A3B-abliterated-v2-4bit
Text Generation • 31B • Updated • 194
The GLM-4 and Z1 series are powerful open-source language models excelling in reasoning, code, and complex tasks.
-
mlx-community/GLM-Z1-32B-0414-4bit
Text Generation • 5B • Updated • 1.82k • 2 -
mlx-community/GLM-4-32B-0414-4bit
Text Generation • 5B • Updated • 2.02k • 4 -
mlx-community/GLM-4-32B-Base-0414-8bit
Text Generation • 9B • Updated • 110 -
mlx-community/GLM-4-32B-Base-0414-6bit
Text Generation • 7B • Updated • 73
Quantization Aware Trained (QAT) Gemma 3 checkpoints. The model preserves similar quality as half precision while using 3x less memory.
-
mlx-community/gemma-3-27b-it-qat-bf16
Image-Text-to-Text • Updated • 2.22k • 4 -
mlx-community/gemma-3-27b-it-qat-8bit
Image-Text-to-Text • Updated • 2.29k • 5 -
mlx-community/gemma-3-27b-it-qat-6bit
Image-Text-to-Text • Updated • 164 -
mlx-community/gemma-3-27b-it-qat-4bit
Image-Text-to-Text • Updated • 182k • 18
-
mlx-community/Llama-4-Scout-17B-16E-Instruct-4bit
Image-Text-to-Text • Updated • 3.56k • 8 -
mlx-community/Llama-4-Scout-17B-16E-Instruct-6bit
Image-Text-to-Text • Updated • 2.49k • 5 -
mlx-community/Llama-4-Scout-17B-16E-Instruct-8bit
Image-Text-to-Text • Updated • 2.6k • 1 -
mlx-community/Llama-4-Maverick-17B-16E-Instruct-4bit
Text Generation • 63B • Updated • 2.79k • 7
-
mlx-community/answerdotai-ModernBERT-base-8bit
Fill-Mask • 0.1B • Updated • 28 -
mlx-community/answerdotai-ModernBERT-base-4bit
Fill-Mask • 0.0B • Updated • 29 -
mlx-community/answerdotai-ModernBERT-base-bf16
Fill-Mask • 0.2B • Updated • 20 -
mlx-community/answerdotai-ModernBERT-Large-Instruct-4bit
Fill-Mask • 0.1B • Updated • 14
A collection of lightweight, state-of-the-art open models built from the same research and technology that powers the Gemini 2.0 models
-
mlx-community/gemma-3-4b-it-8bit
Image-Text-to-Text • 2B • Updated • 9.94k • 3 -
mlx-community/gemma-3-4b-pt-4bit
Image-Text-to-Text • 1B • Updated • 1.84k • 3 -
mlx-community/gemma-3-4b-it-bf16
Image-Text-to-Text • 5B • Updated • 2.4k -
mlx-community/gemma-3-4b-pt-6bit
Image-Text-to-Text • 1B • Updated • 24
-
mlx-community/OLMoE-1B-7B-0125-Instruct
Text Generation • 7B • Updated • 16 -
mlx-community/OLMoE-1B-7B-0125-Instruct-8bit
Text Generation • 2B • Updated • 11 -
mlx-community/OLMoE-1B-7B-0125-Instruct-6bit
Text Generation • 2B • Updated • 11 -
mlx-community/OLMoE-1B-7B-0125-Instruct-4bit
Text Generation • 1B • Updated • 16 • 1
-
mlx-community/olmOCR-7B-0225-preview-bf16
Image-Text-to-Text • 8B • Updated • 106 • 3 -
mlx-community/olmOCR-7B-0225-preview-4bit
Image-Text-to-Text • 2B • Updated • 1.86k • 1 -
mlx-community/olmOCR-7B-0225-preview-6bit
Image-Text-to-Text • 2B • Updated • 38 -
mlx-community/olmOCR-7B-0225-preview-8bit
Updated
Kokoro is an open-weight TTS model with 82 million parameters. Despite its lightweight architecture, it delivers amazing quality.
FuseAI is attempting to merge CoT models to achieve newer models that are more than the sum of their parts.
-
mlx-community/Qwen2.5-VL-72B-Instruct-8bit
Image-Text-to-Text • 21B • Updated • 139 • 2 -
mlx-community/Qwen2.5-VL-72B-Instruct-6bit
Image-Text-to-Text • 16B • Updated • 58 • 1 -
mlx-community/Qwen2.5-VL-72B-Instruct-4bit
Image-Text-to-Text • 12B • Updated • 2.03k • 6 -
mlx-community/Qwen2.5-VL-72B-Instruct-3bit
Image-Text-to-Text • 10B • Updated • 144 • 4
-
mlx-community/Qwen2.5-7B-Instruct-1M-4bit
Text Generation • 1B • Updated • 2.34k • 9 -
mlx-community/Qwen2.5-7B-Instruct-1M-6bit
Text Generation • 2B • Updated • 38 • 2 -
mlx-community/Qwen2.5-7B-Instruct-1M-3bit
Text Generation • 1.0B • Updated • 13 -
mlx-community/Qwen2.5-7B-Instruct-1M-8bit
Text Generation • 2B • Updated • 60 • 3
Convert HTML content to LLM-friendly Markdown/JSON content
Kyutai's Helium-1 2B Model, outperforming other state of the art small models.
-
mlx-community/helium-1-preview-2b-float32
Text Generation • 2B • Updated • 9 -
mlx-community/helium-1-preview-2b
Text Generation • 2B • Updated • 58 -
mlx-community/helium-1-preview-2b-8bit
Text Generation • 0.6B • Updated • 25 • 1 -
mlx-community/helium-1-preview-2b-4bit
Text Generation • 0.3B • Updated • 12 • 1
-
mlx-community/QVQ-72B-Preview-4bit
Image-Text-to-Text • 11B • Updated • 48 • 7 -
mlx-community/QVQ-72B-Preview-6bit
Image-Text-to-Text • 16B • Updated • 14 • 2 -
mlx-community/QVQ-72B-Preview-3bit
Image-Text-to-Text • 9B • Updated • 30 • 5 -
mlx-community/QVQ-72B-Preview-8bit
Image-Text-to-Text • 21B • Updated • 20 • 3
-
mlx-community/deepseek-vl2-6bit
Image-Text-to-Text • 6B • Updated • 31 • 1 -
mlx-community/deepseek-vl2-small-4bit
Image-Text-to-Text • 3B • Updated • 81 -
mlx-community/deepseek-vl2-4bit
Image-Text-to-Text • 4B • Updated • 78 • 1 -
mlx-community/deepseek-vl2-small-6bit
Image-Text-to-Text • 4B • Updated • 32
The best uncensored models
-
mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1
Text Generation • 8B • Updated • 24 -
mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-8bit
Text Generation • 2B • Updated • 40 -
mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-6bit
Text Generation • 2B • Updated • 25 -
mlx-community/Josiefied-Qwen2.5-Coder-7B-Instruct-abliterated-v1-4bit
Text Generation • 1B • Updated • 71 • 1
EXAONE 3.5, a collection of instruction-tuned bilingual generative models ranging from 2.4B to 32B parameters, developed by LG AI.
-
mlx-community/Llama-3.3-70B-Instruct-8bit
Text Generation • 20B • Updated • 2.55k • 12 -
mlx-community/Llama-3.3-70B-Instruct-6bit
Text Generation • 15B • Updated • 215 • 5 -
mlx-community/Llama-3.3-70B-Instruct-3bit
Text Generation • 9B • Updated • 321 • 6 -
mlx-community/Llama-3.3-70B-Instruct-4bit
Text Generation • 11B • Updated • 5.03k • 28
-
mlx-community/paligemma2-3b-ft-docci-448-8bit
Image-Text-to-Text • 0.9B • Updated • 16 -
mlx-community/paligemma2-3b-ft-docci-448-6bit
Image-Text-to-Text • 0.7B • Updated • 41 -
mlx-community/paligemma2-3b-ft-docci-448-bf16
Image-Text-to-Text • 3B • Updated • 18 • 1 -
mlx-community/paligemma2-10b-ft-docci-448-bf16
Image-Text-to-Text • 10B • Updated • 46 • 3
-
mlx-community/SmolVLM-Instruct-4bit
Image-Text-to-Text • 0.5B • Updated • 1.91k • 4 -
mlx-community/SmolVLM-Instruct-6bit
Image-Text-to-Text • 0.6B • Updated • 15 -
mlx-community/SmolVLM-Instruct-8bit
Image-Text-to-Text • 0.7B • Updated • 75 • 9 -
mlx-community/SmolVLM-Instruct-bf16
Image-Text-to-Text • 2B • Updated • 40 • 4
-
mlx-community/Florence-2-base-ft-4bit
Image-Text-to-Text • 0.0B • Updated • 100 • 2 -
mlx-community/Florence-2-large-ft-bf16
Image-Text-to-Text • 0.8B • Updated • 70 • 1 -
mlx-community/Florence-2-base-ft-bf16
Image-Text-to-Text • 0.3B • Updated • 24 • 1 -
mlx-community/Florence-2-base-ft-8bit
Image-Text-to-Text • 0.1B • Updated • 29 • 1
Falcon Mamba models compatible with MLX
Code-specific model series based on Qwen2.5
-
mlx-community/Qwen2.5-Coder-32B-Instruct-8bit
Text Generation • 9B • Updated • 265 • 10 -
mlx-community/Qwen2.5-Coder-14B-Instruct-4bit
Text Generation • 2B • Updated • 2.3k • 4 -
mlx-community/Qwen2.5-Coder-14B-Instruct-bf16
Text Generation • 15B • Updated • 41 • 2 -
mlx-community/Qwen2.5-Coder-3B-Instruct-8bit
Text Generation • 0.9B • Updated • 173
A collection of Neversleep's RP focused Lumimaid LLMs.
Qwen1.5 is the improved version of Qwen, the large language model series developed by Alibaba Cloud.
-
mlx-community/Qwen1.5-1.8B-Chat-4bit
Text Generation • 0.5B • Updated • 50 • 2 -
mlx-community/Qwen1.5-0.5B-Chat-4bit
Text Generation • 0.1B • Updated • 4.21k • 4 -
mlx-community/Qwen1.5-14B-Chat-4bit
Text Generation • 3B • Updated • 23 • 1 -
mlx-community/Qwen1.5-7B-Chat-4bit
Text Generation • 2B • Updated • 118 • 2
Google’s Code-Gemma
-
mlx-community/Meta-Llama-3-8B-Instruct-4bit
Text Generation • 2B • Updated • 10.3k • 78 -
mlx-community/Meta-Llama-3-8B-4bit
Text Generation • 2B • Updated • 113 • 8 -
mlx-community/Meta-Llama-Guard-2-8B-4bit
Text Generation • 2B • Updated • 27 -
mlx-community/Meta-Llama-3-70B-4bit
Text Generation • 11B • Updated • 131 • 9
The Qwen 2.5 models are a series of AI models trained on 18 trillion tokens, supporting 29 languages and offering advanced features such as instructio
-
mlx-community/Qwen2.5-72B-Instruct-bf16
Text Generation • 73B • Updated • 25 -
mlx-community/Qwen2.5-72B-Instruct-8bit
Text Generation • 20B • Updated • 52 • 4 -
mlx-community/Qwen2.5-72B-Instruct-4bit
Text Generation • 11B • Updated • 1.89k • 6 -
mlx-community/Qwen2.5-32B-Instruct-bf16
Text Generation • 33B • Updated • 27
-
mlx-community/Phi-3-mini-4k-instruct-4bit
Text Generation • 0.6B • Updated • 2.08k • 11 -
mlx-community/Phi-3-mini-128k-instruct-4bit
Text Generation • 0.6B • Updated • 1.91k • 12 -
mlx-community/Phi-3-mini-128k-instruct-8bit
Text Generation • 1B • Updated • 75 • 10 -
mlx-community/Phi-3-mini-4k-instruct-8bit
Text Generation • 1B • Updated • 89 • 2
OpenAI Whisper speech recognition models in MLX format
A family of Open-source Efficient Language Models from Apple.
Mamba is a new LLM architecture that integrates the Structured State Space sequence model to manage lengthy data sequences.
A series of smol LLMs: 135M, 360M and 1.7B.
-
mlx-community/Meta-Llama-3.1-70B-bf16
Text Generation • 71B • Updated • 52 • 3 -
mlx-community/Meta-Llama-3.1-70B-Instruct-bf16
Text Generation • 71B • Updated • 35 • 2 -
mlx-community/Meta-Llama-3.1-8B-Instruct-bf16
Text Generation • 8B • Updated • 119 • • 3 -
mlx-community/Meta-Llama-3.1-8B-Instruct-8bit
Text Generation • 2B • Updated • 2.18k • 9
EnCodec models in MLX
Meta goes small with Llama3.2, both text only 1B and 3B, and the 11B Vision models.
-
mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated
Image-Text-to-Text • 11B • Updated • 2.25k • 7 -
mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated-8-bit
Image-Text-to-Text • 3B • Updated • 151 -
mlx-community/Llama-3.2-11B-Vision-Instruct-abliterated-4-bit
Image-Text-to-Text • 2B • Updated • 186 • 1 -
mlx-community/Llama-3.2-11B-Vision-Instruct-8bit
Image-Text-to-Text • 3B • Updated • 2.58k • 10