Jorge Leon

leonsarmiento

14 2 50

azappa's profile picture

Alekstest's profile picture

mikkokam's profile picture

https://benchmark-yourself.streamlit.app/

leonsarmiento

AI & ML interests

AI for Environment

Recent Activity

liked a model 1 day ago

InternScience/Agents-A1-4B

updated a model 7 days ago

leonsarmiento/Occamy-1.0-6bit-XL-mlx

published a model 7 days ago

leonsarmiento/Occamy-1.0-6bit-XL-mlx

View all activity

Organizations

leonsarmiento 's collections 12

Local SOTA for RAM Poor Mac users

High quality low-bit BaseQuant_XL for 16 ~ 32 GB M-base-chip Macs (Mac Mini, MacBook Air, base MacBook Pro): bf16 routing, 8-bit attn, MoE arch.

leonsarmiento/gemma-4-26B-A4B-it-3bit-XL-mlx

Image-Text-to-Text • 26B • Updated 13 days ago • 108
leonsarmiento/Qwen3.8-27B-3bit-mtp-mlx

Image-Text-to-Text • 28B • Updated 13 days ago • 3.2k • 8
leonsarmiento/Qwen3.6-35B-A3B-3bit-XL-mlx

Image-Text-to-Text • 35B • Updated 13 days ago • 74
leonsarmiento/Huihui-Qwen3.6-35B-A3B-abliterated-3bit-XL-mlx

Image-Text-to-Text • 35B • Updated 13 days ago • 357

Local SOTA for 48GB Macs

Best local MLX-quantized LLMs fitting 48GB unified memory. 6-bit XL variants are data-agnostic honest kings; oQ4e-XL for RAM-desperate coders.

leonsarmiento/gemma-4-26B-A4B-it-6bit-XL-mlx

Image-Text-to-Text • 26B • Updated 16 days ago • 230 • 3
leonsarmiento/Qwen3.6-35B-A3B-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 154 • 1
leonsarmiento/Qwen3.6-35B-A3B-oQ4e-mtp-XL-mlx

Image-Text-to-Text • 36B • Updated 13 days ago • 246 • 1
leonsarmiento/Ornith-1.0-35B-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 40

Ornith-Agents-A1-merge

DARE-TIES merge of Ornith-1.0-35B and Agents-A1 (Qwen3.5-35B-A3B MoE) — BaseQuant_XL MLX quantizations

leonsarmiento/Ornith-Agents-A1-3.6-35B-A3B-dare_ties-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 136 • 1
leonsarmiento/Ornith-Agents-A1-3.6-35B-A3B-task_arithmetic_v2-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 145 • 1

Qwen3.6-27B MLX Quantizations

MLX quantizations based on Qwen3.6-27B dense architecture. Includes vanilla and uncensored-heretic variants.

leonsarmiento/Qwen3.6-27B-3bit-mlx

Text Generation • 27B • Updated Jul 22 • 430 • 12
leonsarmiento/Qwen3.6-27B-uncensored-heretic-v2-3bit-mlx

Text Generation • 27B • Updated Jul 22 • 203 • 2
leonsarmiento/ThinkingCap-Qwen3.6-27B-3bit-mlx

Text Generation • 27B • Updated Jul 22 • 95

Ornith-35B-A3B

Ornith-1.0-35B MoE (35B/~3B active). Vanilla + uncensored-heretic. BaseQuant_XL recommended.

leonsarmiento/Ornith-1.0-35B-5bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 231 • 11
leonsarmiento/Ornith-1.0-35B-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 40
leonsarmiento/Ornith-1.0-35B-5bit-mlx

Image-Text-to-Text • 35B • Updated Jul 22 • 151 • 6
leonsarmiento/Ornith-1.0-35B-uncensored-heretic-5bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 414 • 3

Gemma 4 26B A4B MLX Quantizations

Gemma-4-26B-A4B QAT MoE (25.2B/3.8B active). Vanilla + Huihui-abliterated. BaseQuant_XL recommended.

leonsarmiento/gemma-4-26B-A4B-it-6bit-XL-mlx

Image-Text-to-Text • 26B • Updated 16 days ago • 230 • 3
leonsarmiento/gemma-4-26B-A4B-it-8bit-XL-mlx

Image-Text-to-Text • 26B • Updated Aug 4 • 35
leonsarmiento/Huihui-gemma-4-26B-A4B-it-6bit-XL-mlx

Image-Text-to-Text • 26B • Updated Aug 4 • 125
leonsarmiento/Huihui-gemma-4-26B-A4B-it-8bit-XL-mlx

Image-Text-to-Text • 26B • Updated Aug 4 • 82

Qwen3.8-27B MLX Quantizations

3-bit VLM base + 4-bit MTP drafter. Pair both for accelerated instruct decode. reasoning_effort baked to low.

leonsarmiento/Qwen3.8-27B-3bit-mlx

Image-Text-to-Text • 27B • Updated Aug 18 • 1.25k • 1
leonsarmiento/Qwen3.8-27B-3bit-mtp-mlx

Image-Text-to-Text • 28B • Updated 13 days ago • 3.2k • 8

Notable, Niche or just weird

One-off MLX quantizations that didn't fit into a model-family collection.

leonsarmiento/Aella-Qwen3-14B-8bit-mlx

Text Generation • 15B • Updated Nov 27, 2025 • 14
leonsarmiento/Assistant_Pepe_8B-8bit-mlx

Text Generation • 8B • Updated Feb 13 • 18
leonsarmiento/Domyn-Small-v1.0-8bit-mlx

Text Generation • 10B • Updated Jun 26 • 31
leonsarmiento/Llama-3.3-8B-Instruct-128K_Abliterated-8bit-mlx

Text Generation • 8B • Updated Jan 19 • 63

Agents-A1 MLX Quantizations

MLX quantizations of Agents-A1 (Qwen3.5 MoE, ~3B active). BaseQuant_XL variants.

leonsarmiento/Agents-A1-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 33

Qwen3.6-35B-A3B MLX Quantizations

Qwen3.6-35B-A3B MoE (35B/~3B active). Vanilla + Huihui-abliterated + heretic. BaseQuant_XL recommended.

leonsarmiento/Qwen3.6-35B-A3B-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 154 • 1
leonsarmiento/Huihui-Qwen3.6-35B-A3B-abliterated-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 211 • 1
leonsarmiento/Qwen3.6-35B-A3B-oQ4e-mtp-XL-mlx

Image-Text-to-Text • 36B • Updated 13 days ago • 246 • 1

Nex-N2-mini MLX Quantizations

MLX quantizations based on Nex-N2-mini (35B MoE, ~3B active, Qwen3.5). Includes vanilla and Huihui-abliterated variants. BaseQuant_XL recommended.

leonsarmiento/Huihui-Nex-N2-mini-abliterated-5bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 48

GLM-4.7-Flash MLX Quantizations

MLX quantizations of zai-org/GLM-4.7-Flash (31B MoE, ~3B active). BaseQuant_XL recommended over standard variants.

leonsarmiento/GLM-4.7-Flash-6bit-XL-mlx

Text Generation • 30B • Updated Aug 4 • 277 • 1
leonsarmiento/GLM-4.7-Flash-6bit-mlx

Text Generation • 30B • Updated Jul 22 • 110

Local SOTA for RAM Poor Mac users

High quality low-bit BaseQuant_XL for 16 ~ 32 GB M-base-chip Macs (Mac Mini, MacBook Air, base MacBook Pro): bf16 routing, 8-bit attn, MoE arch.

leonsarmiento/gemma-4-26B-A4B-it-3bit-XL-mlx

Image-Text-to-Text • 26B • Updated 13 days ago • 108
leonsarmiento/Qwen3.8-27B-3bit-mtp-mlx

Image-Text-to-Text • 28B • Updated 13 days ago • 3.2k • 8
leonsarmiento/Qwen3.6-35B-A3B-3bit-XL-mlx

Image-Text-to-Text • 35B • Updated 13 days ago • 74
leonsarmiento/Huihui-Qwen3.6-35B-A3B-abliterated-3bit-XL-mlx

Image-Text-to-Text • 35B • Updated 13 days ago • 357

Qwen3.8-27B MLX Quantizations

3-bit VLM base + 4-bit MTP drafter. Pair both for accelerated instruct decode. reasoning_effort baked to low.

leonsarmiento/Qwen3.8-27B-3bit-mlx

Image-Text-to-Text • 27B • Updated Aug 18 • 1.25k • 1
leonsarmiento/Qwen3.8-27B-3bit-mtp-mlx

Image-Text-to-Text • 28B • Updated 13 days ago • 3.2k • 8

Local SOTA for 48GB Macs

Best local MLX-quantized LLMs fitting 48GB unified memory. 6-bit XL variants are data-agnostic honest kings; oQ4e-XL for RAM-desperate coders.

leonsarmiento/gemma-4-26B-A4B-it-6bit-XL-mlx

Image-Text-to-Text • 26B • Updated 16 days ago • 230 • 3
leonsarmiento/Qwen3.6-35B-A3B-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 154 • 1
leonsarmiento/Qwen3.6-35B-A3B-oQ4e-mtp-XL-mlx

Image-Text-to-Text • 36B • Updated 13 days ago • 246 • 1
leonsarmiento/Ornith-1.0-35B-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 40

Notable, Niche or just weird

One-off MLX quantizations that didn't fit into a model-family collection.

leonsarmiento/Aella-Qwen3-14B-8bit-mlx

Text Generation • 15B • Updated Nov 27, 2025 • 14
leonsarmiento/Assistant_Pepe_8B-8bit-mlx

Text Generation • 8B • Updated Feb 13 • 18
leonsarmiento/Domyn-Small-v1.0-8bit-mlx

Text Generation • 10B • Updated Jun 26 • 31
leonsarmiento/Llama-3.3-8B-Instruct-128K_Abliterated-8bit-mlx

Text Generation • 8B • Updated Jan 19 • 63

Ornith-Agents-A1-merge

DARE-TIES merge of Ornith-1.0-35B and Agents-A1 (Qwen3.5-35B-A3B MoE) — BaseQuant_XL MLX quantizations

leonsarmiento/Ornith-Agents-A1-3.6-35B-A3B-dare_ties-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 136 • 1
leonsarmiento/Ornith-Agents-A1-3.6-35B-A3B-task_arithmetic_v2-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 145 • 1

Agents-A1 MLX Quantizations

MLX quantizations of Agents-A1 (Qwen3.5 MoE, ~3B active). BaseQuant_XL variants.

leonsarmiento/Agents-A1-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 33

Qwen3.6-27B MLX Quantizations

MLX quantizations based on Qwen3.6-27B dense architecture. Includes vanilla and uncensored-heretic variants.

leonsarmiento/Qwen3.6-27B-3bit-mlx

Text Generation • 27B • Updated Jul 22 • 430 • 12
leonsarmiento/Qwen3.6-27B-uncensored-heretic-v2-3bit-mlx

Text Generation • 27B • Updated Jul 22 • 203 • 2
leonsarmiento/ThinkingCap-Qwen3.6-27B-3bit-mlx

Text Generation • 27B • Updated Jul 22 • 95

Qwen3.6-35B-A3B MLX Quantizations

Qwen3.6-35B-A3B MoE (35B/~3B active). Vanilla + Huihui-abliterated + heretic. BaseQuant_XL recommended.

leonsarmiento/Qwen3.6-35B-A3B-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 154 • 1
leonsarmiento/Huihui-Qwen3.6-35B-A3B-abliterated-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 211 • 1
leonsarmiento/Qwen3.6-35B-A3B-oQ4e-mtp-XL-mlx

Image-Text-to-Text • 36B • Updated 13 days ago • 246 • 1

Ornith-35B-A3B

Ornith-1.0-35B MoE (35B/~3B active). Vanilla + uncensored-heretic. BaseQuant_XL recommended.

leonsarmiento/Ornith-1.0-35B-5bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 231 • 11
leonsarmiento/Ornith-1.0-35B-6bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 40
leonsarmiento/Ornith-1.0-35B-5bit-mlx

Image-Text-to-Text • 35B • Updated Jul 22 • 151 • 6
leonsarmiento/Ornith-1.0-35B-uncensored-heretic-5bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 414 • 3

Nex-N2-mini MLX Quantizations

MLX quantizations based on Nex-N2-mini (35B MoE, ~3B active, Qwen3.5). Includes vanilla and Huihui-abliterated variants. BaseQuant_XL recommended.

leonsarmiento/Huihui-Nex-N2-mini-abliterated-5bit-XL-mlx

Image-Text-to-Text • 35B • Updated Aug 4 • 48

Gemma 4 26B A4B MLX Quantizations

Gemma-4-26B-A4B QAT MoE (25.2B/3.8B active). Vanilla + Huihui-abliterated. BaseQuant_XL recommended.

leonsarmiento/gemma-4-26B-A4B-it-6bit-XL-mlx

Image-Text-to-Text • 26B • Updated 16 days ago • 230 • 3
leonsarmiento/gemma-4-26B-A4B-it-8bit-XL-mlx

Image-Text-to-Text • 26B • Updated Aug 4 • 35
leonsarmiento/Huihui-gemma-4-26B-A4B-it-6bit-XL-mlx

Image-Text-to-Text • 26B • Updated Aug 4 • 125
leonsarmiento/Huihui-gemma-4-26B-A4B-it-8bit-XL-mlx

Image-Text-to-Text • 26B • Updated Aug 4 • 82

GLM-4.7-Flash MLX Quantizations

MLX quantizations of zai-org/GLM-4.7-Flash (31B MoE, ~3B active). BaseQuant_XL recommended over standard variants.

leonsarmiento/GLM-4.7-Flash-6bit-XL-mlx

Text Generation • 30B • Updated Aug 4 • 277 • 1
leonsarmiento/GLM-4.7-Flash-6bit-mlx

Text Generation • 30B • Updated Jul 22 • 110