Table of Contents
📦 Loading Models
A model can be loaded by its Model ID: 
var model = LM.LoadFromModelID("gemma3:4b");

LM-Kit Model Catalog

Filters 0
Showing 0 models
Model & ID ⇅ Capabilities Context ⇅ Params. ⇅ Format ⇅ License ⇅ Download Details
BAAI bge m3bge-m3 Text Embeddings 8,192 0.57 B GGUF mit bge-m3-Q4_K_M.gguf Details →
BAAI bge m3 reranker v2bge-m3-reranker Text Reranking 8,192 0.57 B GGUF apache-2.0 Bge-M3-568M-Q4_K_M.gguf Details →
BAAI bge small en v1.5bge-small Text Embeddings 512 0.03 B GGUF mit bge-small-en-v1.5-f16.gguf Details →
DeepSeek Coder V2 Litedeepseek-coder-v2:16b Code Completion 163,840 15.71 B GGUF deepseek DeepSeek-Coder-2-Lite-15.7B-Instruct-Q4_K_M.gguf Details →
DeepSeek R1 Distill Llamadeepseek-r1:8b Text Generation Chat Code Completion Math 131,072 8.03 B GGUF mit DeepSeek-R1-Distill-Llama-8B-Q4_K_M.gguf Details →
Google Gemma Embedding 300Membeddinggemma-300m Text Embeddings 2,058 0.03 B GGUF gemma embeddinggemma-300M-Q4_K_M.gguf Details →
TII Falcon 3 Instructfalcon3:3b Text Generation Chat Code Completion Math 32,768 3.23 B GGUF falcon-llm-license Falcon3-3B-Instruct-q4_k_m.gguf Details →
TII Falcon 3 Instructfalcon3:7b Text Generation Chat Code Completion Math 32,768 7.62 B GGUF falcon-llm-license Falcon-3-7.6B-Instruct-Q4_K_M.gguf Details →
TII Falcon 3 Instructfalcon3:10b Text Generation Chat Code Completion Math 32,768 10.31 B GGUF falcon-llm-license Falcon3-10B-Instruct-q4_k_m.gguf Details →
Google Gemma 2gemma2:2b Text Generation Chat 8,192 2.61 B GGUF gemma gemma-2-2B-Q4_K_M.gguf Replaced by
gemma3:1b
Google Gemma 2gemma2:9b Text Generation Chat 8,192 9.24 B GGUF gemma gemma-2-9B-Q4_K_M.gguf Replaced by
gemma3:4b
Google Gemma 2gemma2:27b Text Generation Chat 8,192 27.23 B GGUF gemma gemma-2-27B-Q4_K_M.gguf Replaced by
gemma3:27b
Google Gemma 3gemma3:1b Text Generation Chat 32,768 1.00 B GGUF gemma gemma-3-it-1B-Q4_K_M.gguf Details →
Google Gemma 3gemma3:4b Text Generation Chat Code Completion Math Vision 131,072 3.88 B GGUF gemma gemma-3-4b-it-Q4_K_M.lmk Details →
Google Gemma 3gemma3:12b Text Generation Chat Code Completion Math Vision 131,072 11.77 B GGUF gemma gemma-3-12b-it-Q4_K_M.lmk Details →
Google Gemma 3gemma3:27b Text Generation Chat Code Completion Math Vision 131,072 27.01 B GGUF gemma gemma-3-27b-it-Q4_K_M.lmk Details →
Google Gemma 3 270Mgemma3:270m Text Generation Chat 32,768 0.27 B GGUF gemma gemma-3-270M-it-Q4_K_M.gguf Details →
OpenAI Gpt OSSgptoss:20b Text Generation Chat Code Completion Math Reasoning Tools Call 131,072 20.91 B GGUF apache-2.0 gpt-oss-20b-mxfp4.gguf Details →
IBM Granite 3.1 Dense Instructgranite3.1-dense:2b Text Generation Chat Code Completion 131,072 2.53 B GGUF apache-2.0 granite-3.1-2.5B-Q4_K_M.gguf Replaced by
granite4-h:3b
IBM Granite 3.1 Dense Instructgranite3.1-dense:8b Text Generation Chat Code Completion 131,072 8.17 B GGUF apache-2.0 granite-3.1-8.2B-Q4_K_M.gguf Replaced by
granite4-h:7b
IBM Granite 3.3 Instructgranite3.3:2b Text Generation Chat Code Completion 131,072 2.53 B GGUF apache-2.0 granite-3.3-8B-Instruct-Q4_K_M.gguf Replaced by
granite4-h:3b
IBM Granite 3.3 Instructgranite3.3:8b Text Generation Chat Code Completion 131,072 8.17 B GGUF apache-2.0 granite-3.3-2B-Instruct-Q4_K_M.gguf Replaced by
granite4-h:7b
IBM Granite 4 Micro Instructgranite4-h:3b Text Generation Chat Code Completion Tools Call 1,048,576 3.19 B GGUF apache-2.0 Granite-4.0-H-Micro-3.2B-Q4_K_M.gguf Details →
IBM Granite 4 Tiny Instructgranite4-h:7b Text Generation Chat Code Completion Tools Call 1,048,576 6.94 B GGUF apache-2.0 Granite-4.0-H-Tiny-64x994M-Q4_K_M.gguf Details →
LightOn LightOnOCR 1025lightonocr1025:1b Text Generation Chat 8,192 1.16 B GGUF apache-2.0 lightonocr-1b-1025-Q4_K_M.lmk Details →
Meta Llama 3.1 Instructllama3.1 Text Generation Chat Tools Call 131,072 8.03 B GGUF llama3.1 Llama-3.1-8B-Instruct-Q4_K_M.gguf Details →
Meta Llama 3.2 Instructllama3.2:1b Text Generation Chat 131,072 1.24 B GGUF llama3.2 Llama-3.2-1B-Instruct-Q4_K_M.gguf Details →
Meta Llama 3.2 Instructllama3.2:3b Text Generation Chat 131,072 3.21 B GGUF llama3.2 Llama-3.2-3B-Instruct-Q4_K_M.gguf Details →
Meta Llama 3.3 Instructllama3.3 Text Generation Chat Code Completion Math Tools Call 131,072 70.55 B GGUF llama3.3 Llama-3.3-70B-Instruct-Q4_K_M.gguf Details →
LM-Kit Sarcasm Detection V1lmkit-sarcasm-detection Sentiment Analysis 2,048 1.10 B GGUF lm-kit LM-Kit.Sarcasm_Detection-TinyLlama-1.1B-1T-OpenOrca-en-q4.gguf Details →
LM-Kit Sentiment Analysis V2lmkit-sentiment-analysis Sentiment Analysis 131,072 1.24 B GGUF lm-kit lm-kit-sentiment-analysis-2.0-1b-q4.gguf Details →
LM-Kit Tasks Previewlmkit-tasks:4b-preview Text Generation Chat Code Completion Math Vision 131,072 3.88 B LMK lmkit lmkit-tasks-4b-preview.lmk Details →
Mistral Magistral Small 1.1magistral-small Text Generation Chat Code Completion Math Reasoning 40,960 23.57 B GGUF apache-2.0 Magistral-Small-2506-Q4_K_M.gguf Replaced by
magistral-small1.2
Mistral Magistral Small 1.2magistral-small1.2 Text Generation Chat Math Reasoning Tools Call 40,960 23.57 B GGUF apache-2.0 Magistral-Small-2509-Q4_K_M.gguf Details →
OpenBMB MiniCPM o 2.6 Visionminicpm-o Text Generation Chat Vision 32,768 8.12 B LMK OpenBMB MiniCPM-o-V-2.6-Q4_K_M.lmk Details →
OpenBMB MiniCPM 2.6 Visionminicpm-v Text Generation Chat Vision 32,768 8.12 B LMK OpenBMB MiniCPM-V-2.6-Q4_K_M.lmk Replaced by
minicpm-o
OpenBMB MiniCPM-V 4.5minicpm-v-45 Text Generation Chat Vision 40,960 8.72 B LMK OpenBMB minicpm-v-4.5-8b.lmk Details →
Mistral Ministral 3ministral3:3b Text Generation Chat Math Vision Tools Call 262,144 3.85 B LMK apache-2.0 ministral-3-3b-instruct-Q4_K_M.lmk Details →
Mistral Ministral 3ministral3:8b Text Generation Chat Math Vision Tools Call 262,144 8.92 B LMK apache-2.0 ministral-3-8b-instruct-Q4_K_M.lmk Details →
Mistral Ministral 3ministral3:14b Text Generation Chat Math Vision Tools Call 262,144 13.95 B LMK apache-2.0 ministral-3-14b-instruct-Q4_K_M.lmk Details →
Mistral Nemo Instruct 2407mistral-nemo Text Generation Chat 1,024,000 12.25 B GGUF apache-2.0 Mistral-Nemo-2407-12.2B-Instruct-Q4_K_M.gguf Replaced by
ministral3:8b
Mistral Small Instruct 2501mistral-small Text Generation Chat Code Completion Math 32,768 23.57 B GGUF apache-2.0 Mistral-Small-Instruct-2501-24B-Q4_K_M.gguf Replaced by
mistral-small3.2
Mistral Small 3.1 Instruct 2503mistral-small3.1 Text Generation Chat Code Completion Math 131,072 23.57 B GGUF apache-2.0 Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf Replaced by
mistral-small3.2
Mistral Small 3.2 Instruct 2503mistral-small3.2 Text Generation Chat Code Completion Math Tools Call 131,072 23.57 B GGUF apache-2.0 Mistral-Small-3.2-24B-Instruct-2506-Q4_K_M.gguf Details →
Nomic embed text v1.5nomic-embed-text Text Embeddings 2,048 0.14 B GGUF apache-2.0 nomic-embed-text-1.5-Q4_K_M.gguf Details →
Nomic embed vision v1.5nomic-embed-vision Image Embeddings 197 0.09 B ONNX apache-2.0 nomic-embed-vision-1.5-Q8.lmk Details →
Microsoft Phi 3.5 Mini Instructphi3.5 Text Generation Chat 131,072 3.82 B GGUF mit Phi-3.5-mini-Instruct-Q4_K_M.gguf Replaced by
phi4-mini
Microsoft Phi 4 Instructphi4 Text Generation Chat Math Tools Call 16,384 14.66 B GGUF mit Phi-4-14.7B-Instruct-Q4_K_M.gguf Details →
Microsoft Phi 4 Mini Instructphi4-mini Text Generation Chat Tools Call 131,072 3.84 B GGUF mit Phi-4-mini-Instruct-Q4_K_M.gguf Details →
Mistral Pixtralpixtral Text Generation Chat Vision 1,024,000 12.68 B LMK apache-2.0 pixtral-12B-Q4_K_M.lmk Details →
Alibaba Qwen 2 Vision Instructqwen2-vl:2b Text Generation Chat Vision 32,768 2.21 B LMK apache-2.0 Qwen2-VL-2B-Instruct-Q4_K_M.lmk Replaced by
qwen3-vl:2b
Alibaba Qwen 2 Vision Instructqwen2-vl:8b Text Generation Chat Vision 32,768 8.29 B LMK apache-2.0 Qwen2-VL-8.3B-Instruct-Q4_K_M.lmk Replaced by
qwen3-vl:8b
Alibaba Qwen 2.5 Instructqwen2.5:0.5b Text Generation Chat 32,768 0.49 B GGUF apache-2.0 Qwen-2.5-0.5B-Instruct-Q4_K_M.gguf Replaced by
qwen3:0.6b
Alibaba Qwen 2.5 Instructqwen2.5:3b Text Generation Chat 32,768 3.09 B GGUF qwen-research Qwen-2.5-3.1B-Instruct-Q4_K_M.gguf Replaced by
qwen3:4b
Alibaba Qwen 2.5 Instructqwen2.5:7b Text Generation Chat 32,768 7.62 B GGUF apache-2.0 Qwen-2.5-7B-Instruct-Q4_K_M.gguf Replaced by
qwen3:8b
Alibaba Qwen 2.5 Vision Instructqwen2.5-vl:3b Text Generation Chat Vision 128,000 3.75 B LMK qwen research license Qwen2.5-VL-3B-Instruct-Q4_K_M.lmk Replaced by
qwen3-vl:4b
Alibaba Qwen 2.5 Vision Instructqwen2.5-vl:7b Text Generation Chat Vision 128,000 8.29 B LMK apache-2.0 Qwen2.5-VL-7B-Instruct-Q4_K_M.lmk Replaced by
qwen3-vl:8b
Alibaba Qwen 2.5 Vision Instructqwen2.5-vl:32b Text Generation Chat Vision 128,000 33.45 B LMK apache-2.0 Qwen2.5-VL-32B-Instruct-Q4_K_M.lmk Details →
Alibaba Qwen 3 Instructqwen3:0.6b Text Generation Chat 40,960 0.75 B GGUF apache-2.0 Qwen3-0.6B-Q4_K_M.gguf Details →
Alibaba Qwen 3 Instructqwen3:1.7b Text Generation Chat 40,960 2.03 B GGUF apache-2.0 Qwen3-1.7B-Q4_K_M.gguf Details →
Alibaba Qwen 3 Instructqwen3:4b Text Generation Chat Math Reasoning Tools Call 40,960 4.02 B GGUF apache-2.0 Qwen3-4B-Q4_K_M.gguf Details →
Alibaba Qwen 3 Instructqwen3:8b Text Generation Chat Math Reasoning Tools Call 40,960 8.19 B GGUF apache-2.0 Qwen3-8B-Q4_K_M.gguf Details →
Alibaba Qwen 3 Instructqwen3:14b Text Generation Chat Math Reasoning Tools Call 40,960 14.77 B GGUF apache-2.0 Qwen3-14B-Q4_K_M.gguf Details →
Alibaba Qwen 3 Embeddingqwen3-embedding:0.6b Text Embeddings 32,768 0.60 B GGUF apache-2.0 Qwen3-Embedding-0.6B-Q4_K_M.gguf Details →
Alibaba Qwen 3 Embeddingqwen3-embedding:4b Text Embeddings 40,960 4.02 B GGUF apache-2.0 Qwen3-Embedding-4B-Q4_K_M.gguf Details →
Alibaba Qwen 3 Embeddingqwen3-embedding:8b Text Embeddings 40,960 7.57 B GGUF apache-2.0 Qwen3-Embedding-8B-Q4_K_M.gguf Details →
Alibaba Qwen 3 Vision Instructqwen3-vl:2b Text Generation Chat Code Completion Math Vision Tools Call 262,144 2.13 B LMK apache-2.0 qwen3-vl-2b-instruct-Q4_K_M.lmk Details →
Alibaba Qwen 3 Vision Instructqwen3-vl:4b Text Generation Chat Code Completion Math Vision Tools Call 262,144 2.13 B LMK apache-2.0 qwen3-vl-4b-instruct-Q4_K_M.lmk Details →
Alibaba Qwen 3 Vision Instructqwen3-vl:8b Text Generation Chat Code Completion Math Vision Tools Call 262,144 8.77 B LMK apache-2.0 qwen3-vl-8b-instruct-Q4_K_M.lmk Details →
Alibaba Qwen 3 Vision Instructqwen3-vl:30b Text Generation Chat Code Completion Math Vision Tools Call 262,144 31.07 B LMK apache-2.0 qwen3-vl-30b-instruct-Q4_K_M.lmk Details →
Alibaba Qwen QwQqwq Text Generation Chat Math Reasoning Tools Call 40,960 32.76 B GGUF apache-2.0 QwQ-32B-Q4_K_M.gguf Details →
HuggingFace SmolLM3smollm3:3b Text Generation Chat Code Completion Math 65,536 3.08 B GGUF apache-2.0 SmolLM3-3B-Q4_K_M.gguf Details →
U2-Net 44Mu2net Image Segmentation 0 0.04 B LMK apache-2.0 u2-net-F32.lmk Details →
OpenAI Whisper Basewhisper-base Speech-to-Text 1,500 0.07 B GGML mit whisper-base-q8_0.bin Details →
OpenAI Whisper Large Turbo V3whisper-large-turbo3 Speech-to-Text 1,500 0.81 B GGML mit whisper-large-v3-turbo-q8_0.bin Details →
OpenAI Whisper Large V3whisper-large3 Speech-to-Text 1,500 1.54 B GGML mit whisper-large-v3-q8_0.bin Details →
OpenAI Whisper Mediumwhisper-medium Speech-to-Text 1,500 0.76 B GGML mit whisper-medium-q8_0.bin Details →
OpenAI Whisper Smallwhisper-small Speech-to-Text 1,500 0.24 B GGML mit whisper-small-q8_0.bin Details →
OpenAI Whisper Tinywhisper-tiny Speech-to-Text 1,500 0.04 B GGML mit whisper-tiny-q8_0.bin Details →

Model Details

BAAI bge m3bge-m3

Description: A unified, multilingual embedding model that delivers dense, sparse, and multi-vector retrieval on texts from short queries up to 8,192-token documents in over 100 languages.

Specifications:

  • Capabilities: Text Embeddings
  • Architecture: bert
  • Context Length: 8,192 tokens
  • Parameter Count: 566,703,104
  • Quantization Precision: 4-bit
  • File Size: 417.50 MB
  • Format: GGUF
  • License: mit
  • SHA256: e251234fcb7d050991a6be491952f485bf5c641dd10c3272dc1301fd281ad50f
⬇️ Download bge-m3-Q4_K_M.gguf
BAAI bge m3 reranker v2bge-m3-reranker

Description: A unified, multilingual reranker that ingests query–document pairs and directly produces sigmoid-normalized relevance scores across over 100 languages.

Specifications:

  • Capabilities: Text Reranking
  • Architecture: bert
  • Context Length: 8,192 tokens
  • Parameter Count: 567,753,729
  • Quantization Precision: 4-bit
  • File Size: 418.07 MB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: ce947cece730cbf7d836da8c5490a9987ef0f919014b9275e7ce9aa12d96e6d9
⬇️ Download Bge-M3-568M-Q4_K_M.gguf
BAAI bge small en v1.5bge-small

Description: An efficient, CPU-friendly English embedding model (BAAI General Embedding) designed for lightweight applications.

Specifications:

  • Capabilities: Text Embeddings
  • Architecture: bert
  • Context Length: 512 tokens
  • Parameter Count: 33,212,160
  • Quantization Precision: 16-bit
  • File Size: 64.45 MB
  • Format: GGUF
  • License: mit
  • SHA256: cd5790da23df71e7e20fe20bb523bd4586a533a4ee813cc562e32b37929141c1
⬇️ Download bge-small-en-v1.5-f16.gguf
DeepSeek Coder V2 Lite 15.7Bdeepseek-coder-v2:16b

Description: An open-source mixture-of-experts code model tailored for code completion tasks. Early evaluations indicated competitive performance relative to leading code models.

Specifications:

  • Capabilities: Code Completion
  • Architecture: deepseek2
  • Context Length: 163,840 tokens
  • Parameter Count: 15,706,484,224
  • Quantization Precision: 4-bit
  • File Size: 9.65 GB
  • Format: GGUF
  • License: deepseek
  • SHA256: ac398e8c1c670d3c362d3c1182614916bab7c364708ec073fcf947f6802d509e
⬇️ Download DeepSeek-Coder-2-Lite-15.7B-Instruct-Q4_K_M.gguf
DeepSeek R1 Distill Llama 8Bdeepseek-r1:8b

Description: DeepSeek-R1 enhances its predecessor by integrating cold-start data to overcome repetition and readability issues, achieving state-of-the-art performance in math, code, and reasoning tasks, with all models open-sourced.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math
  • Architecture: llama
  • Context Length: 131,072 tokens
  • Parameter Count: 8,030,261,312
  • Quantization Precision: 4-bit
  • File Size: 4.58 GB
  • Format: GGUF
  • License: mit
  • SHA256: 596fce705423e44831fe63367a30ccc7b36921c1bfdd4b9dfde85a5aa97ac2ef
⬇️ Download DeepSeek-R1-Distill-Llama-8B-Q4_K_M.gguf
Google Gemma Embedding 300Membeddinggemma-300m

Description: EmbeddingGemma 300M is an open, state-of-the-art-for-its-size embedding model from Google DeepMind (Gemma 3, T5Gemma-initialized). It produces 768-dimensional text vectors for search/retrieval, classification, clustering, and semantic similarity across 100+ languages, and supports Matryoshka Representation Learning (truncate to 512/256/128 with re-normalization). Optimized for on-device/CPU deployment.

Specifications:

  • Capabilities: Text Embeddings
  • Architecture: gemma-embedding
  • Context Length: 2,058 tokens
  • Parameter Count: 33,212,160
  • Quantization Precision: 4-bit
  • File Size: 288.83 MB
  • Format: GGUF
  • License: gemma
  • SHA256: 3d55e7fe66eb4c7b2d01b4fbd30c00dc7a101bd6c9f724a6e7e5cfaa87968420
⬇️ Download embeddinggemma-300M-Q4_K_M.gguf
TII Falcon 3 Instruct 3.2Bfalcon3:3b

Description: Designed for multilingual tasks including chat, text generation, and code completion, supporting extended context lengths.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math
  • Architecture: llama
  • Context Length: 32,768 tokens
  • Parameter Count: 3,227,655,168
  • Quantization Precision: 4-bit
  • File Size: 1.87 GB
  • Format: GGUF
  • License: falcon-llm-license
  • SHA256: 81c6b52d221c2f0eea3db172fc74de28534f2fd15f198ecbfcc55577d20cbf8a
⬇️ Download Falcon3-3B-Instruct-q4_k_m.gguf
TII Falcon 3 Instruct 7.6Bfalcon3:7b

Description: Offers robust performance across chat, text generation, and mathematical reasoning tasks with extended context support.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math
  • Architecture: llama
  • Context Length: 32,768 tokens
  • Parameter Count: 7,615,616,512
  • Quantization Precision: 4-bit
  • File Size: 4.26 GB
  • Format: GGUF
  • License: falcon-llm-license
  • SHA256: 4ce1da546d76e04ce77eb076556eb25e1096faf6155ee429245e4bfa3f5ddf5d
⬇️ Download Falcon-3-7.6B-Instruct-Q4_K_M.gguf
TII Falcon 3 Instruct 10.3Bfalcon3:10b

Description: A larger variant tailored for multilingual dialogue, code completion, and complex reasoning tasks with extended context support.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math
  • Architecture: llama
  • Context Length: 32,768 tokens
  • Parameter Count: 10,305,653,760
  • Quantization Precision: 4-bit
  • File Size: 5.86 GB
  • Format: GGUF
  • License: falcon-llm-license
  • SHA256: a0c0edbd35019ff26d972a0373b25b4c8d72315395a3b6036aca5e6bafa3d819
⬇️ Download Falcon3-10B-Instruct-q4_k_m.gguf
Google Gemma 2 2.6Bgemma2:2b

Description: A lightweight decoder-only model from Google, available in both pre-trained and instruction-tuned variants for text-to-text tasks.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: gemma2
  • Context Length: 8,192 tokens
  • Parameter Count: 2,614,341,888
  • Quantization Precision: 4-bit
  • File Size: 1.59 GB
  • Format: GGUF
  • License: gemma
  • SHA256: 362d09c1496e035ecf0737d8fe03e8e607c61e57e16b22cedd158525f6721e06

⚠️ Replaced by: gemma3:1b

⬇️ Download gemma-2-2B-Q4_K_M.gguf
Google Gemma 2 9.2Bgemma2:9b

Description: A decoder-only text-to-text model from Google, offering competitive performance in both pre-trained and instruction-tuned configurations.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: gemma2
  • Context Length: 8,192 tokens
  • Parameter Count: 9,241,705,984
  • Quantization Precision: 4-bit
  • File Size: 5.37 GB
  • Format: GGUF
  • License: gemma
  • SHA256: b6059a960d2f4f881630f1e795b40f7e09e5e12d3a6b1900474d6108ea880afd

⚠️ Replaced by: gemma3:4b

⬇️ Download gemma-2-9B-Q4_K_M.gguf
Google Gemma 2 27.2Bgemma2:27b

Description: A larger variant in the Gemma 2 family, optimized for text generation and instruction following with open weights provided.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: gemma2
  • Context Length: 8,192 tokens
  • Parameter Count: 27,227,128,320
  • Quantization Precision: 4-bit
  • File Size: 15.50 GB
  • Format: GGUF
  • License: gemma
  • SHA256: bb4b276745da743d550720dc2e6c847498eef45e7b82a4d5a73ef6636f78027a

⚠️ Replaced by: gemma3:27b

⬇️ Download gemma-2-27B-Q4_K_M.gguf
Google Gemma 3 1Bgemma3:1b

Description: Gemma is Google's lightweight, multimodal, open AI model family based on Gemini technology, supporting text and image inputs, 128K context windows, multilingual capabilities in over 140 languages, and optimized for resource-limited environments.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: gemma3
  • Context Length: 32,768 tokens
  • Parameter Count: 999,885,952
  • Quantization Precision: 4-bit
  • File Size: 768.72 MB
  • Format: GGUF
  • License: gemma
  • SHA256: bacfe3de6eee9fba412d5c0415630172c2a602dae26bb353e1b20dd67194a226
⬇️ Download gemma-3-it-1B-Q4_K_M.gguf
Google Gemma 3 3.9Bgemma3:4b

Description: Gemma is Google's lightweight, multimodal, open AI model family based on Gemini technology, supporting text and image inputs, 128K context windows, multilingual capabilities in over 140 languages, and optimized for resource-limited environments.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Vision
  • Architecture: gemma3
  • Context Length: 131,072 tokens
  • Parameter Count: 3,880,099,328
  • Quantization Precision: 4-bit
  • File Size: 2.87 GB
  • Format: GGUF
  • License: gemma
  • SHA256: abb283e96c0abf58468a18127ce6e8b2bfc98e48f1ec618f658495c09254bdae
⬇️ Download gemma-3-4b-it-Q4_K_M.lmk
Google Gemma 3 11.8Bgemma3:12b

Description: Gemma is Google's lightweight, multimodal, open AI model family based on Gemini technology, supporting text and image inputs, 128K context windows, multilingual capabilities in over 140 languages, and optimized for resource-limited environments.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Vision
  • Architecture: gemma3
  • Context Length: 131,072 tokens
  • Parameter Count: 11,765,788,416
  • Quantization Precision: 4-bit
  • File Size: 7.35 GB
  • Format: GGUF
  • License: gemma
  • SHA256: d6f01cdb4369769ea87c5211a7fd865e12dbb9e2a937b43ef281a5b7e9ba2e35
⬇️ Download gemma-3-12b-it-Q4_K_M.lmk
Google Gemma 3 27.2Bgemma3:27b

Description: Gemma is Google's lightweight, multimodal, open AI model family based on Gemini technology, supporting text and image inputs, 128K context windows, multilingual capabilities in over 140 languages, and optimized for resource-limited environments.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Vision
  • Architecture: gemma3
  • Context Length: 131,072 tokens
  • Parameter Count: 27,009,002,240
  • Quantization Precision: 4-bit
  • File Size: 15.97 GB
  • Format: GGUF
  • License: gemma
  • SHA256: 2d0e4382259ae2da28b9c0342e982a58eafbddad7c05bbfe6e104f2b3c165994
⬇️ Download gemma-3-27b-it-Q4_K_M.lmk
Google Gemma 3 270Mgemma3:270m

Description: Gemma is Google's lightweight, multimodal, open AI model family based on Gemini technology, supporting text and image inputs, 128K context windows, multilingual capabilities in over 140 languages, and optimized for resource-limited environments.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: gemma3
  • Context Length: 32,768 tokens
  • Parameter Count: 268,098,176
  • Quantization Precision: 4-bit
  • File Size: 241.39 MB
  • Format: GGUF
  • License: gemma
  • SHA256: e28b323bc75925d6edc8d3f030268830bf53c59c296d77278ac24653403d9d47
⬇️ Download gemma-3-270M-it-Q4_K_M.gguf
OpenAI Gpt OSSgptoss:20b

Description: OpenAI’s medium-sized open-weight Mixture-of-Experts model (≈21B params; ~3.6B active per token). This MXFP4 GGUF build is optimized for local inference, supports long context (131k), strong reasoning & tool use, and can run on consumer GPUs (~16GB VRAM).

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Reasoning Tools Call
  • Architecture: gpt-oss
  • Context Length: 131,072 tokens
  • Parameter Count: 20,914,757,184
  • Quantization Precision: 4-bit
  • File Size: 11.28 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 52f57ab7d3df3ba9173827c1c6832e73375553a846f3e32b49f1ae2daad688d4
⬇️ Download gpt-oss-20b-mxfp4.gguf
IBM Granite 3.1 Dense Instruct 2.5Bgranite3.1-dense:2b

Description: A long-context instruct model finetuned with a mix of open source and synthetic datasets. Designed for dialogue and text generation tasks.

Specifications:

  • Capabilities: Text Generation Chat Code Completion
  • Architecture: granite
  • Context Length: 131,072 tokens
  • Parameter Count: 2,533,531,648
  • Quantization Precision: 4-bit
  • File Size: 1.44 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: ba05b36d0a8cebf8ccd13bbbb904bebe182f4854fbcff19cd1ee54bc82bbd298

⚠️ Replaced by: granite4-h:3b

⬇️ Download granite-3.1-2.5B-Q4_K_M.gguf
IBM Granite 3.1 Dense Instruct 8.2Bgranite3.1-dense:8b

Description: An extended-context model optimized for dialogue and code completion tasks. Developed with diverse training data to enhance long-context understanding.

Specifications:

  • Capabilities: Text Generation Chat Code Completion
  • Architecture: granite
  • Context Length: 131,072 tokens
  • Parameter Count: 8,170,848,256
  • Quantization Precision: 4-bit
  • File Size: 4.60 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: d1ada98d7b274fc6b119bd19b8d3536cd006544e9aae06db6f8b2c256d584044

⚠️ Replaced by: granite4-h:7b

⬇️ Download granite-3.1-8.2B-Q4_K_M.gguf
IBM Granite 3.3 Instruct 2.5Bgranite3.3:2b

Description: A long-context instruct model finetuned with a mix of open source and synthetic datasets. Designed for dialogue and text generation tasks.

Specifications:

  • Capabilities: Text Generation Chat Code Completion
  • Architecture: granite
  • Context Length: 131,072 tokens
  • Parameter Count: 2,533,539,840
  • Quantization Precision: 4-bit
  • File Size: 1.44 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: dbe4dd51bd6c1e39f96c831bf086454c9b313bd1c279ebb7166f2a37d86598da

⚠️ Replaced by: granite4-h:3b

⬇️ Download granite-3.3-8B-Instruct-Q4_K_M.gguf
IBM Granite 3.3 Instruct 8.2Bgranite3.3:8b

Description: An extended-context model optimized for dialogue and code completion tasks. Developed with diverse training data to enhance long-context understanding.

Specifications:

  • Capabilities: Text Generation Chat Code Completion
  • Architecture: granite
  • Context Length: 131,072 tokens
  • Parameter Count: 8,170,864,640
  • Quantization Precision: 4-bit
  • File Size: 4.60 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 1c890e740d7ecb010716a858eda315c01ac5bb0edfaf68bf17118868a26bb8ff

⚠️ Replaced by: granite4-h:7b

⬇️ Download granite-3.3-2B-Instruct-Q4_K_M.gguf
IBM Granite 4 Micro Instruct 3.2Bgranite4-h:3b

Description: Hybrid long-context instruct model (Mamba-2 + attention) finetuned from Granite-4.0-H-Micro-Base with SFT, RL alignment, and model merging. Delivers stronger instruction following and robust tool/function calling in multilingual dialog (en, de, es, fr, ja, pt, ar, cs, it, ko, nl, zh), with 1M-token context for enterprise assistants. Excels at summarization, classification, extraction, QA/RAG, and code—including FIM—and supports structured chat templates and OpenAI-style tool schemas.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Tools Call
  • Architecture: granitehybrid
  • Context Length: 1,048,576 tokens
  • Parameter Count: 3,191,396,096
  • Quantization Precision: 4-bit
  • File Size: 1.81 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: dbe7b747aa49340f80629811652636b55f4ca4cbbb92ee7e17c442d8a1130566
⬇️ Download Granite-4.0-H-Micro-3.2B-Q4_K_M.gguf
IBM Granite 4 Tiny Instruct 6.9Bgranite4-h:7b

Description: Granite-4.0-H-Tiny is an ~7B-parameter hybrid (attention + Mamba2) MoE decoder with a 1M-token context, instruction-tuned (SFT, RL alignment, model merging) for enterprise assistants. It improves instruction following and tool/function calling, supports multilingual dialog (en, de, es, fr, ja, pt, ar, cs, it, ko, nl, zh), and excels at summarization, classification, extraction, QA/RAG, and code/FIM tasks.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Tools Call
  • Architecture: granitehybrid
  • Context Length: 1,048,576 tokens
  • Parameter Count: 6,939,037,248
  • Quantization Precision: 4-bit
  • File Size: 3.94 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 75234a50a38235dd4c891dae1a702ccf47a5d89da751d38f90a43be4794f18fb
⬇️ Download Granite-4.0-H-Tiny-64x994M-Q4_K_M.gguf
LightOn LightOnOCR 1025 1Blightonocr1025:1b

Description: LightOnOCR-1B-1025 is a compact, end-to-end vision–language model for high-accuracy OCR and document understanding, delivering fast, layout-aware text extraction from complex documents.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: qwen3
  • Context Length: 8,192 tokens
  • Parameter Count: 1,161,230,336
  • Quantization Precision: 4-bit
  • File Size: 710.36 MB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 42646c9408fd59f342dec352bfc3c84c2255b03e79500506ee9d320e2f66be37
⬇️ Download lightonocr-1b-1025-Q4_K_M.lmk
Meta Llama 3.1 Instruct 8Bllama3.1

Description: A multilingual generative model optimized for dialogue and text generation tasks. Designed for robust performance on common benchmarks.

Specifications:

  • Capabilities: Text Generation Chat Tools Call
  • Architecture: llama
  • Context Length: 131,072 tokens
  • Parameter Count: 8,030,261,312
  • Quantization Precision: 4-bit
  • File Size: 4.58 GB
  • Format: GGUF
  • License: llama3.1
  • SHA256: ad00fe50a62d1e009b4e06cd57ab55c9a30cbf5e7f183de09115d75ada73bd5b
⬇️ Download Llama-3.1-8B-Instruct-Q4_K_M.gguf
Meta Llama 3.2 Instruct 1.2Bllama3.2:1b

Description: A multilingual instruct-tuned model optimized for dialogue, retrieval, and summarization tasks.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: llama
  • Context Length: 131,072 tokens
  • Parameter Count: 1,235,814,432
  • Quantization Precision: 4-bit
  • File Size: 770.28 MB
  • Format: GGUF
  • License: llama3.2
  • SHA256: 88725e821cf35f1a0dbeaa4a3bebeb91e6c6b6a9d50f808ab42d64233284cce1
⬇️ Download Llama-3.2-1B-Instruct-Q4_K_M.gguf
Meta Llama 3.2 Instruct 3.2Bllama3.2:3b

Description: A multilingual dialogue model with robust text generation and summarization capabilities.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: llama
  • Context Length: 131,072 tokens
  • Parameter Count: 3,212,749,888
  • Quantization Precision: 4-bit
  • File Size: 1.88 GB
  • Format: GGUF
  • License: llama3.2
  • SHA256: 6810bf3cce69d440a22b85a3b3e28f57c868f1c98686abd995f1dc5d9b955cfe
⬇️ Download Llama-3.2-3B-Instruct-Q4_K_M.gguf
Meta Llama 3.3 Instruct 70.6Bllama3.3

Description: A large multilingual generative model optimized for dialogue, text tasks, code completion, and mathematical reasoning with extended context support.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Tools Call
  • Architecture: llama
  • Context Length: 131,072 tokens
  • Parameter Count: 70,553,706,560
  • Quantization Precision: 4-bit
  • File Size: 39.60 GB
  • Format: GGUF
  • License: llama3.3
  • SHA256: 57f78fe3b141afa56406278265656524c51c9837edb3537ad43708b6d4ecc04d
⬇️ Download Llama-3.3-70B-Instruct-Q4_K_M.gguf
LM-Kit Sarcasm Detection V1 1.1Blmkit-sarcasm-detection

Description: Optimized for detecting sarcasm in English text within the LM-Kit framework. Suitable for CPU-based inference.

Specifications:

  • Capabilities: Sentiment Analysis
  • Architecture: llama
  • Context Length: 2,048 tokens
  • Parameter Count: 1,100,048,384
  • Quantization Precision: 4-bit
  • File Size: 636.88 MB
  • Format: GGUF
  • License: lm-kit
  • SHA256: cc82abd224dba9c689b19d368db6078d6167ca84897b21870d7d6a2c0f09d7d0
⬇️ Download LM-Kit.Sarcasm_Detection-TinyLlama-1.1B-1T-OpenOrca-en-q4.gguf
LM-Kit Sentiment Analysis V2 1.2Blmkit-sentiment-analysis

Description: Designed for multilingual sentiment analysis tasks, this LM-Kit model is optimized for efficient CPU-based inference.

Specifications:

  • Capabilities: Sentiment Analysis
  • Architecture: llama
  • Context Length: 131,072 tokens
  • Parameter Count: 1,235,814,432
  • Quantization Precision: 4-bit
  • File Size: 770.28 MB
  • Format: GGUF
  • License: lm-kit
  • SHA256: e12f4abf6453a8431985ce1d6350c265cd58b25210156a917e3608c850fd7add
⬇️ Download lm-kit-sentiment-analysis-2.0-1b-q4.gguf
LM-Kit Tasks 4B Previewlmkit-tasks:4b-preview

Description: A 4B-parameter Gemma3-based model optimized for LM-Kit tasks. Achieves state-of-the-art performance in classification, structured data extraction, language detection, and sentiment analysis, while also supporting chat, embeddings, text generation, code completion, math reasoning, and vision understanding. Designed for seamless integration into LM-Kit pipelines to deliver efficient, reliable, and high-quality results across domains.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Vision
  • Architecture: gemma3
  • Context Length: 131,072 tokens
  • Parameter Count: 3,880,099,328
  • Quantization Precision: 4-bit
  • File Size: 3.09 GB
  • Format: LMK
  • License: lmkit
  • SHA256: 3ec9fe4622e2d9a050b3d2c7d2244a911aab75372b04a7bc30bb72a05bdd645c
⬇️ Download lmkit-tasks-4b-preview.lmk
Mistral Magistral Small 1.1 24Bmagistral-small

Description: Building upon Mistral Small 3.1 (2503), with added reasoning capabilities, undergoing SFT from Magistral Medium traces and RL on top, it's a small, efficient reasoning model with 24B parameters.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Reasoning
  • Architecture: llama
  • Context Length: 40,960 tokens
  • Parameter Count: 23,572,403,200
  • Quantization Precision: 4-bit
  • File Size: 13.35 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 7680ba6895d405340f1461cb835a147055689a37d88b193cc5a365aaea76da9e

⚠️ Replaced by: magistral-small1.2

⬇️ Download Magistral-Small-2506-Q4_K_M.gguf
Mistral Magistral Small 1.2 24Bmagistral-small1.2

Description: Magistral Small 1.2 (2509) builds upon Mistral Small 3.2 (2506) with added reasoning via SFT from Magistral Medium traces and RL, special [THINK]/[/THINK] tokens, and a 128K context. This GGUF release is text-only (no vision encoder) and should be paired with mistral-common for the correct chat template.

Specifications:

  • Capabilities: Text Generation Chat Math Reasoning Tools Call
  • Architecture: llama
  • Context Length: 40,960 tokens
  • Parameter Count: 23,572,403,200
  • Quantization Precision: 4-bit
  • File Size: 13.35 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: d3a024d29e0e8f35d9353f5d4f08fc3715406835c8ae1328ce2f3bd212a43434
⬇️ Download Magistral-Small-2509-Q4_K_M.gguf
OpenBMB MiniCPM o 2.6 Vision 8.1Bminicpm-o

Description: An end-to-end multimodal model supporting real-time speech, image, and text understanding. Offers enhanced performance for conversational tasks.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: qwen2
  • Context Length: 32,768 tokens
  • Parameter Count: 8,116,736,752
  • Quantization Precision: 4-bit
  • File Size: 5.00 GB
  • Format: LMK
  • License: OpenBMB
  • SHA256: 6fd17ed1f46bfcddb5a3482dd882dd022a46aa8c33cb93d75f809cd4d118ab53
⬇️ Download MiniCPM-o-V-2.6-Q4_K_M.lmk
OpenBMB MiniCPM 2.6 Vision 8.1Bminicpm-v

Description: A multimodal model designed for vision and text tasks, built upon SigLip and Qwen architectures. Evaluate performance against current benchmarks.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: qwen2
  • Context Length: 32,768 tokens
  • Parameter Count: 8,116,736,752
  • Quantization Precision: 4-bit
  • File Size: 5.00 GB
  • Format: LMK
  • License: OpenBMB
  • SHA256: a10b1aa434899ea0bd5bb5e281f622fed0b02434241d53435fce05773fa7cfa8

⚠️ Replaced by: minicpm-o

⬇️ Download MiniCPM-V-2.6-Q4_K_M.lmk
OpenBMB MiniCPM-V 4.5 8Bminicpm-v-45

Description: MiniCPM-V 4.5 is a state-of-the-art multimodal LLM built on Qwen3-8B and SigLIP2-400M. It delivers GPT-4o-level performance for single-image, multi-image, and high-FPS video understanding on local devices. The model supports controllable fast/deep thinking, real-time speech and text comprehension, strong OCR and document parsing (up to 1.8M pixels), and multilingual capabilities in 30+ languages. Optimized for efficiency, it enables CPU inference, mobile deployment, and scalable usage through formats like LMK, GGUF, and AWQ.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: qwen3
  • Context Length: 40,960 tokens
  • Parameter Count: 8,715,965,680
  • Quantization Precision: 4-bit
  • File Size: 5.70 GB
  • Format: LMK
  • License: OpenBMB
  • SHA256: 000c56809f033e53637f364461cfadb8c4aa09e533a3fde66de39cbb41bf5cb7
⬇️ Download minicpm-v-4.5-8b.lmk
Mistral Ministral 3 3Bministral3:3b

Description: Smallest member of the Ministral 3 family, an edge-optimized multilingual instruct model with a 256K context window and solid reasoning and code capabilities for constrained hardware.

Specifications:

  • Capabilities: Text Generation Chat Math Vision Tools Call
  • Architecture: mistral3
  • Context Length: 262,144 tokens
  • Parameter Count: 3,849,093,120
  • Quantization Precision: 4-bit
  • File Size: 2.42 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 300ce4373c14a3c1d68e37a7f4537b98776eb8ccac50ff32b30cc8ae6191ee96
⬇️ Download ministral-3-3b-instruct-Q4_K_M.lmk
Mistral Ministral 3 8Bministral3:8b

Description: Mid-sized Ministral 3 variant that balances quality and cost, offering strong multilingual reasoning, math, and code performance with a 256K context while remaining practical for single-GPU and edge deployments.

Specifications:

  • Capabilities: Text Generation Chat Math Vision Tools Call
  • Architecture: mistral3
  • Context Length: 262,144 tokens
  • Parameter Count: 8,918,030,336
  • Quantization Precision: 4-bit
  • File Size: 5.27 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 4bd03de58774150e6d19e9a1f0e4f1e010784ac7b98801c35144f5fd796c81c8
⬇️ Download ministral-3-8b-instruct-Q4_K_M.lmk
Mistral Ministral 3 14Bministral3:14b

Description: Flagship Ministral 3 instruct model delivering frontier-level multilingual reasoning, math, and code performance in a 256K-token context, with design tuned for efficient edge and single-GPU deployment.

Specifications:

  • Capabilities: Text Generation Chat Math Vision Tools Call
  • Architecture: mistral3
  • Context Length: 262,144 tokens
  • Parameter Count: 13,945,036,800
  • Quantization Precision: 4-bit
  • File Size: 8.11 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 032759d8b1814347ae571a2dd6d24c2d36760141f07756a6559bb77a17a9e821
⬇️ Download ministral-3-14b-instruct-Q4_K_M.lmk
Mistral Nemo Instruct 2407 12.2Bmistral-nemo

Description: An instruct-tuned variant developed in collaboration with NVIDIA, balancing model size with performance for conversational tasks.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: llama
  • Context Length: 1,024,000 tokens
  • Parameter Count: 12,247,782,400
  • Quantization Precision: 4-bit
  • File Size: 6.96 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 579ab8f5178f5900d0c4e14534929aa0dba97e3f97be76b31ebe537ffd6cf169

⚠️ Replaced by: ministral3:8b

⬇️ Download Mistral-Nemo-2407-12.2B-Instruct-Q4_K_M.gguf
Mistral Small Instruct 2501 24Bmistral-small

Description: Optimized for local deployment, this model balances parameter count and performance for chat and code tasks.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math
  • Architecture: llama
  • Context Length: 32,768 tokens
  • Parameter Count: 23,572,403,200
  • Quantization Precision: 4-bit
  • File Size: 13.35 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 4395b5c6136e29e9b11bdba2ee189302ad45dd5c3ef45073b729f077b8f0cec8

⚠️ Replaced by: mistral-small3.2

⬇️ Download Mistral-Small-Instruct-2501-24B-Q4_K_M.gguf
Mistral Small 3.1 Instruct 2503 24Bmistral-small3.1

Description: Mistral Small 3.1 (24B) enhances Mistral Small 3 with advanced vision, 128k context, multilingual support, agentic features, and efficient local deployment.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math
  • Architecture: llama
  • Context Length: 131,072 tokens
  • Parameter Count: 23,572,403,200
  • Quantization Precision: 4-bit
  • File Size: 13.35 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 68922ff3a311c81bc4e983f86e665a12213ee84710c210522f10e65ce980bda7

⚠️ Replaced by: mistral-small3.2

⬇️ Download Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf
Mistral Small 3.2 Instruct 2503 24Bmistral-small3.2

Description: Mistral Small 3.2 (24B) enhances Mistral Small 3 with advanced vision, 128k context, multilingual support, agentic features, and efficient local deployment.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Tools Call
  • Architecture: llama
  • Context Length: 131,072 tokens
  • Parameter Count: 23,572,403,200
  • Quantization Precision: 4-bit
  • File Size: 13.35 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: e1e9a516a90387ec98bb9c45c37dbb1478008d1fa46b216cca893cf008d92c29
⬇️ Download Mistral-Small-3.2-24B-Instruct-2506-Q4_K_M.gguf
Nomic embed text v1.5nomic-embed-text

Description: Provides flexible production embeddings using Matryoshka Representation Learning.

Specifications:

  • Capabilities: Text Embeddings
  • Architecture: nomic-bert
  • Context Length: 2,048 tokens
  • Parameter Count: 136,731,648
  • Quantization Precision: 4-bit
  • File Size: 85.86 MB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 1a60949a331b30bb754ad60b7bdff80d8e563a56b3f7f3f1aed68db8c143003e
⬇️ Download nomic-embed-text-1.5-Q4_K_M.gguf
Nomic embed vision v1.5nomic-embed-vision

Description: ViT-B/16-based image embedding model trained on 1.5B image-text pairs using Matryoshka Representation Learning. Outputs 768-dim embeddings aligned with Nomic Embed Text v1.5 for multimodal search, retrieval, and zero-shot classification.

Specifications:

  • Capabilities: Image Embeddings
  • Architecture: ViT-B/16
  • Context Length: 197 tokens
  • Parameter Count: 92,384,769
  • Quantization Precision: 8-bit
  • File Size: 92.26 MB
  • Format: ONNX
  • License: apache-2.0
  • SHA256: 4f6f6a765625a4b74ec3e62141b7b83e1db1fb904afeda1fa00c1fefefbcc714
⬇️ Download nomic-embed-vision-1.5-Q8.lmk
Microsoft Phi 3.5 Mini Instruct 3.8Bphi3.5

Description: A lightweight model optimized for reasoning-dense tasks and extended context support. Designed for efficient instruction following.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: phi3
  • Context Length: 131,072 tokens
  • Parameter Count: 3,821,079,648
  • Quantization Precision: 4-bit
  • File Size: 2.23 GB
  • Format: GGUF
  • License: mit
  • SHA256: 782c34ae79564d1d92bd44dec233182559b3ecf6fedee44417e2a28c89bd0721

⚠️ Replaced by: phi4-mini

⬇️ Download Phi-3.5-mini-Instruct-Q4_K_M.gguf
Microsoft Phi 4 Instruct 14.7Bphi4

Description: An enhanced generative model trained on a diverse dataset to improve instruction adherence and reasoning capabilities.

Specifications:

  • Capabilities: Text Generation Chat Math Tools Call
  • Architecture: phi3
  • Context Length: 16,384 tokens
  • Parameter Count: 14,659,507,200
  • Quantization Precision: 4-bit
  • File Size: 8.43 GB
  • Format: GGUF
  • License: mit
  • SHA256: 03af8f5c5a87d526047f5c20c99e32bbafd5db6dbfdee8d498d0fe1a3c45af55
⬇️ Download Phi-4-14.7B-Instruct-Q4_K_M.gguf
Microsoft Phi 4 Mini Instruct 3.8Bphi4-mini

Description: A lightweight open model from the Phi-4 family that uses synthetic and curated public data for reasoning-dense outputs, supports a 128K token context, and is enhanced through fine-tuning and preference optimization for precise instruction adherence and robust safety.

Specifications:

  • Capabilities: Text Generation Chat Tools Call
  • Architecture: phi3
  • Context Length: 131,072 tokens
  • Parameter Count: 3,836,021,856
  • Quantization Precision: 4-bit
  • File Size: 2.32 GB
  • Format: GGUF
  • License: mit
  • SHA256: 556492e72efc8d33406b236830ad38d25669482ea7ad91fc643de237e942b9f9
⬇️ Download Phi-4-mini-Instruct-Q4_K_M.gguf
Mistral Pixtral 12Bpixtral

Description: Pixtral 12B is a natively multimodal model combining a 12 B parameter decoder with a 400 M vision encoder, trained on interleaved image–text data for variable image sizes, offering state-of-the-art performance in its weight class across multimodal and text-only benchmarks and supporting ultra-long 128 k sequence lengths.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: llama
  • Context Length: 1,024,000 tokens
  • Parameter Count: 12,682,744,832
  • Quantization Precision: 4-bit
  • File Size: 7.39 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 28d42e60b5f33765ac6f3882abc4c7fd9f5a7955910ff117c13dbfc5aa6bf159
⬇️ Download pixtral-12B-Q4_K_M.lmk
Alibaba Qwen 2 Vision Instruct 2.2Bqwen2-vl:2b

Description: A multilingual vision-language model featuring dynamic resolution processing for advanced image and long-video understanding.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: qwen2vl
  • Context Length: 32,768 tokens
  • Parameter Count: 2,208,985,700
  • Quantization Precision: 4-bit
  • File Size: 1.27 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: b4e546acfd2271f5a0960b64445cae1091e5fc4192d74db72ae57c28729bd0b8

⚠️ Replaced by: qwen3-vl:2b

⬇️ Download Qwen2-VL-2B-Instruct-Q4_K_M.lmk
Alibaba Qwen 2 Vision Instruct 8.3Bqwen2-vl:8b

Description: An extended variant in the Qwen 2 Vision family for multilingual vision-language tasks, including advanced video analysis.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: qwen2vl
  • Context Length: 32,768 tokens
  • Parameter Count: 8,291,375,716
  • Quantization Precision: 4-bit
  • File Size: 4.72 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 90b3eb60611559ba7521590ecccdf1d2a4dfab007566221c6a42f19b91b48686

⚠️ Replaced by: qwen3-vl:8b

⬇️ Download Qwen2-VL-8.3B-Instruct-Q4_K_M.lmk
Alibaba Qwen 2.5 Instruct 0.5Bqwen2.5:0.5b

Description: A compact variant from the Alibaba Qwen 2.5 family, optimized for instruction following across chat, embeddings, and text generation tasks.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: qwen2
  • Context Length: 32,768 tokens
  • Parameter Count: 494,032,768
  • Quantization Precision: 4-bit
  • File Size: 379.38 MB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 09b44ff0ef0a160ffe50778c0828754201bb3a40522a941839c23acfbc9ceec0

⚠️ Replaced by: qwen3:0.6b

⬇️ Download Qwen-2.5-0.5B-Instruct-Q4_K_M.gguf
Alibaba Qwen 2.5 Instruct 3.1Bqwen2.5:3b

Description: A mid-sized model from the Alibaba Qwen 2.5 series, designed for diverse tasks including chat, embeddings, and text generation. Performance should be evaluated relative to current benchmarks.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: qwen2
  • Context Length: 32,768 tokens
  • Parameter Count: 3,085,938,688
  • Quantization Precision: 4-bit
  • File Size: 1.80 GB
  • Format: GGUF
  • License: qwen-research
  • SHA256: fb88cca2303e7f7d4d52679d633efe66d9c3e3555573b4444abe5ab8af4a97f7

⚠️ Replaced by: qwen3:4b

⬇️ Download Qwen-2.5-3.1B-Instruct-Q4_K_M.gguf
Alibaba Qwen 2.5 Instruct 7.6Bqwen2.5:7b

Description: A larger variant from the Alibaba Qwen 2.5 series that supports extended context and multiple tasks including chat, embeddings, and text generation.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: qwen2
  • Context Length: 32,768 tokens
  • Parameter Count: 7,615,616,512
  • Quantization Precision: 4-bit
  • File Size: 4.36 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 2bf11b8a7d566bddfcc2b222ed7b918afc51239c5f919532de8b9403981ad866

⚠️ Replaced by: qwen3:8b

⬇️ Download Qwen-2.5-7B-Instruct-Q4_K_M.gguf
Alibaba Qwen 2.5 Vision Instruct 3Bqwen2.5-vl:3b

Description: Qwen2.5 VL 3B Instruct is a compact vision-language chat model that delivers advanced object and text/chart understanding, agentic tool-driven interactions, long-video event localization, precise visual grounding with JSON outputs, and structured data extraction, powered by an optimized ViT encoder with dynamic temporal training.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: qwen2vl
  • Context Length: 128,000 tokens
  • Parameter Count: 3,754,622,976
  • Quantization Precision: 4-bit
  • File Size: 2.58 GB
  • Format: LMK
  • License: qwen research license
  • SHA256: 78fee4fde9f7fd93e1365cae46668184a259b1bd2a3169915a4a1e7495f859f8

⚠️ Replaced by: qwen3-vl:4b

⬇️ Download Qwen2.5-VL-3B-Instruct-Q4_K_M.lmk
Alibaba Qwen 2.5 Vision Instruct 7Bqwen2.5-vl:7b

Description: Qwen2.5 VL 7B Instruct is a next-generation vision-language chat model that combines advanced object and text/chart understanding, agentic tool use, long-video event localization, precise visual grounding with JSON outputs, and structured data extraction, all powered by a streamlined ViT encoder with dynamic temporal training.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: qwen2vl
  • Context Length: 128,000 tokens
  • Parameter Count: 8,292,166,656
  • Quantization Precision: 4-bit
  • File Size: 5.16 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: e9a99c7bb06c23bd60594cebf8a881af13f502742df3047eaa3b466c747f7453

⚠️ Replaced by: qwen3-vl:8b

⬇️ Download Qwen2.5-VL-7B-Instruct-Q4_K_M.lmk
Alibaba Qwen 2.5 Vision Instruct 32Bqwen2.5-vl:32b

Description: Qwen2.5 VL 32B Instruct is a next-generation vision-language chat model that combines advanced object and text/chart understanding, agentic tool use, long-video event localization, precise visual grounding with JSON outputs, and structured data extraction, all powered by a streamlined ViT encoder with dynamic temporal training.

Specifications:

  • Capabilities: Text Generation Chat Vision
  • Architecture: qwen2vl
  • Context Length: 128,000 tokens
  • Parameter Count: 33,452,718,336
  • Quantization Precision: 4-bit
  • File Size: 5.16 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 9081e05fc2832177162b9ea8ccde1e0fdb1d8ed429a838527af36de966e2fb92
⬇️ Download Qwen2.5-VL-32B-Instruct-Q4_K_M.lmk
Alibaba Qwen 3 Instruct 0.6Bqwen3:0.6b

Description: Qwen3 is the latest generation of Qwen large language models, combining dense and MoE architectures with seamless “thinking” vs. “non‐thinking” mode switching to deliver state-of-the-art reasoning, coding, agent integration, and instruction-following across 100+ languages.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: qwen3
  • Context Length: 40,960 tokens
  • Parameter Count: 751,632,384
  • Quantization Precision: 4-bit
  • File Size: 461.79 MB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 2b1a7ed56061ad1275847412f61e8e009ada37ef865dccc25747dcc76eea9811
⬇️ Download Qwen3-0.6B-Q4_K_M.gguf
Alibaba Qwen 3 Instruct 1.7Bqwen3:1.7b

Description: Qwen3 is the latest generation of Qwen large language models, combining dense and MoE architectures with seamless “thinking” vs. “non‐thinking” mode switching to deliver state-of-the-art reasoning, coding, agent integration, and instruction-following across 100+ languages.

Specifications:

  • Capabilities: Text Generation Chat
  • Architecture: qwen3
  • Context Length: 40,960 tokens
  • Parameter Count: 2,031,739,904
  • Quantization Precision: 4-bit
  • File Size: 1.19 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: b047d6617eba56dcfa3357566b06807f54b15816faf6182aabd12d7e2378e537
⬇️ Download Qwen3-1.7B-Q4_K_M.gguf
Alibaba Qwen 3 Instruct 4Bqwen3:4b

Description: Qwen3 is the latest generation of Qwen large language models, combining dense and MoE architectures with seamless “thinking” vs. “non‐thinking” mode switching to deliver state-of-the-art reasoning, coding, agent integration, and instruction-following across 100+ languages.

Specifications:

  • Capabilities: Text Generation Chat Math Reasoning Tools Call
  • Architecture: qwen3
  • Context Length: 40,960 tokens
  • Parameter Count: 4,022,468,096
  • Quantization Precision: 4-bit
  • File Size: 2.33 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 9dbc1e801f001ea316a627bb867fdd192fc3b36046fd69e160155ddc12129dbe
⬇️ Download Qwen3-4B-Q4_K_M.gguf
Alibaba Qwen 3 Instruct 8Bqwen3:8b

Description: Qwen3 is the latest generation of Qwen large language models, combining dense and MoE architectures with seamless “thinking” vs. “non‐thinking” mode switching to deliver state-of-the-art reasoning, coding, agent integration, and instruction-following across 100+ languages.

Specifications:

  • Capabilities: Text Generation Chat Math Reasoning Tools Call
  • Architecture: qwen3
  • Context Length: 40,960 tokens
  • Parameter Count: 8,190,735,360
  • Quantization Precision: 4-bit
  • File Size: 4.68 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: b9059e3978453f50a8e9e45a825243abdb8739b2f4623e541fd5a392d9672c0f
⬇️ Download Qwen3-8B-Q4_K_M.gguf
Alibaba Qwen 3 Instruct 14Bqwen3:14b

Description: Qwen3 is the latest generation of Qwen large language models, combining dense and MoE architectures with seamless “thinking” vs. “non‐thinking” mode switching to deliver state-of-the-art reasoning, coding, agent integration, and instruction-following across 100+ languages.

Specifications:

  • Capabilities: Text Generation Chat Math Reasoning Tools Call
  • Architecture: qwen3
  • Context Length: 40,960 tokens
  • Parameter Count: 14,768,307,200
  • Quantization Precision: 4-bit
  • File Size: 8.38 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 520369028ee99e4a3ca413a35126337038a8da561927f81322c1b34aed10e03d
⬇️ Download Qwen3-14B-Q4_K_M.gguf
Alibaba Qwen 3 Embedding 0.6Bqwen3-embedding:0.6b

Description: Lightweight member of the Qwen3 Embedding series, optimized for fast, low-resource semantic search and ranking while preserving strong multilingual and long-context understanding.

Specifications:

  • Capabilities: Text Embeddings
  • Architecture: qwen3
  • Context Length: 32,768 tokens
  • Parameter Count: 595,776,512
  • Quantization Precision: 4-bit
  • File Size: 609.54 MB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: b624c62027986bc4181eadcad0cee479916c498d1039f7063195fd4c14803023
⬇️ Download Qwen3-Embedding-0.6B-Q4_K_M.gguf
Alibaba Qwen 3 Embedding 4Bqwen3-embedding:4b

Description: Mid-size Qwen3 Embedding model offering a strong accuracy–efficiency trade-off for multilingual retrieval, reranking, classification, clustering, and bitext mining with instruction-aware embeddings.

Specifications:

  • Capabilities: Text Embeddings
  • Architecture: qwen3
  • Context Length: 40,960 tokens
  • Parameter Count: 4,021,774,336
  • Quantization Precision: 4-bit
  • File Size: 2.33 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: ac48f080498db5874835a0c6db52aa8a726f8f88fca1dbbb26fc51d5311acb85
⬇️ Download Qwen3-Embedding-4B-Q4_K_M.gguf
Alibaba Qwen 3 Embedding 8Bqwen3-embedding:8b

Description: Flagship Qwen3 Embedding model delivering state-of-the-art multilingual and cross-lingual embeddings for dense retrieval, reranking, text/code search, clustering, and classification, with flexible output dimensions.

Specifications:

  • Capabilities: Text Embeddings
  • Architecture: qwen3
  • Context Length: 40,960 tokens
  • Parameter Count: 7,567,295,488
  • Quantization Precision: 4-bit
  • File Size: 4.36 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 3822cc88f2f3e9a08c4b9bae87261a7e94c503fe0372ad9b6c5b80161886291a
⬇️ Download Qwen3-Embedding-8B-Q4_K_M.gguf
Alibaba Qwen 3 Vision Instruct 2Bqwen3-vl:2b

Description: Qwen3-VL is the latest Qwen vision-language family with stronger text understanding/generation, deeper visual reasoning, native 256K context (expandable), upgraded OCR (32 langs), long-video comprehension, and agentic GUI/tool use. The 2B Instruct edition targets edge devices for multimodal chat, grounding, and document/image understanding.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Vision Tools Call
  • Architecture: qwen3vl
  • Context Length: 262,144 tokens
  • Parameter Count: 2,127,532,032
  • Quantization Precision: 4-bit
  • File Size: 1.29 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 296c414827f80a0205371f2d407169f71457c4b9f9ed49edd1b28e8b9f697ace
⬇️ Download qwen3-vl-2b-instruct-Q4_K_M.lmk
Alibaba Qwen 3 Vision Instruct 4Bqwen3-vl:4b

Description: Qwen3-VL 4B Instruct balances efficiency and quality with advanced spatial perception (2D/3D grounding), timestamp-aligned video reasoning, and “visual coding” (HTML/CSS/JS from images). Suited for on-device or small-GPU multimodal assistants, retrieval, and structured understanding.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Vision Tools Call
  • Architecture: qwen3vl
  • Context Length: 262,144 tokens
  • Parameter Count: 2,127,532,032
  • Quantization Precision: 4-bit
  • File Size: 2.59 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 30c7ff027b6b148950533fbc6cf473abc288245af61cd23ca3ae135b6db9f3e8
⬇️ Download qwen3-vl-4b-instruct-Q4_K_M.lmk
Alibaba Qwen 3 Vision Instruct 8Bqwen3-vl:8b

Description: Qwen3-VL 8B Instruct is the highest-quality dense variant, delivering state-of-the-art multimodal reasoning, long-horizon video understanding, stronger recognition (celebrities, products, flora/fauna, etc.), and robust agent/tool interaction—ideal for high-fidelity VLM chat and STEM tasks.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Vision Tools Call
  • Architecture: qwen3vl
  • Context Length: 262,144 tokens
  • Parameter Count: 8,767,123,696
  • Quantization Precision: 4-bit
  • File Size: 5.21 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 76031476ae3f8720e01a941e655981c7b6779b5709a746879d771534a3d0ccdf
⬇️ Download qwen3-vl-8b-instruct-Q4_K_M.lmk
Alibaba Qwen 3 Vision Instruct 30Bqwen3-vl:30b

Description: Qwen3-VL 30B Instruct is the flagship MoE vision-language model in the Qwen3 family, delivering state-of-the-art multimodal reasoning, long-horizon video and document understanding with a native 256K context, precise spatial grounding and timestamp-aligned event localization, expanded OCR in 32 languages, and powerful visual-agent capabilities for GUI control, visual coding, and tool-augmented STEM and math workflows.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math Vision Tools Call
  • Architecture: qwen3vlmoe
  • Context Length: 262,144 tokens
  • Parameter Count: 31,070,754,032
  • Quantization Precision: 4-bit
  • File Size: 17.79 GB
  • Format: LMK
  • License: apache-2.0
  • SHA256: 955041de8ecc63327b6a5fa408ccf03e80e62e663b88f5314aa2901c332b7478
⬇️ Download qwen3-vl-30b-instruct-Q4_K_M.lmk
Alibaba Qwen QwQ 32.5Bqwq

Description: QwQ is a reasoning-focused model in the Qwen series that significantly outperforms conventional instruction-tuned models on challenging tasks, with QwQ-32B demonstrating competitive performance compared to top reasoning models like DeepSeek-R1 and o1-mini.

Specifications:

  • Capabilities: Text Generation Chat Math Reasoning Tools Call
  • Architecture: qwen2
  • Context Length: 40,960 tokens
  • Parameter Count: 32,763,876,352
  • Quantization Precision: 4-bit
  • File Size: 18.49 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 6c2c72d16bbf5b0c30ac22031e0800b982b7d5c4e4d27daa62b66ee61c565d17
⬇️ Download QwQ-32B-Q4_K_M.gguf
HuggingFace SmolLM3 3Bsmollm3:3b

Description: SmolLM3 is a 3B parameter language model designed to push the boundaries of small models. It supports 6 languages, advanced reasoning and long context. SmolLM3 is a fully open model that offers strong performance at the 3B–4B scale.

Specifications:

  • Capabilities: Text Generation Chat Code Completion Math
  • Architecture: smollm3
  • Context Length: 65,536 tokens
  • Parameter Count: 3,075,098,624
  • Quantization Precision: 4-bit
  • File Size: 1.78 GB
  • Format: GGUF
  • License: apache-2.0
  • SHA256: 03cc959aceff388ca737e4714a20cf4b3ef403116c9759a2a99f504dec40294e
⬇️ Download SmolLM3-3B-Q4_K_M.gguf
U2-Net 44Mu2net

Description: A U-square nested U-Net for salient object detection and general image segmentation; lightweight encoder–decoder with RSU blocks.

Specifications:

  • Capabilities: Image Segmentation
  • Architecture: u2net
  • Context Length: 0 tokens
  • Parameter Count: 44,000,000
  • Quantization Precision: 32-bit
  • File Size: 167.85 MB
  • Format: LMK
  • License: apache-2.0
  • SHA256: bfc5e34225e3c8d3b5c3ffd3b128c7d7e6bb17de9bde56b3a6d0654de5e73661
⬇️ Download u2-net-F32.lmk
OpenAI Whisper Basewhisper-base

Description: A balanced Whisper model delivering moderate resource use with reliable transcription accuracy.

Specifications:

  • Capabilities: Speech-to-Text
  • Architecture: whisper
  • Context Length: 1,500 tokens
  • Parameter Count: 72,593,920
  • Quantization Precision: 8-bit
  • File Size: 77.98 MB
  • Format: GGML
  • License: mit
  • SHA256: c577b9a86e7e048a0b7eada054f4dd79a56bbfa911fbdacf900ac5b567cbb7d9
⬇️ Download whisper-base-q8_0.bin
OpenAI Whisper Large Turbo V3whisper-large-turbo3

Description: A turbo-optimized Whisper large v3 variant for faster transcription with near-v3 accuracy.

Specifications:

  • Capabilities: Speech-to-Text
  • Architecture: whisper
  • Context Length: 1,500 tokens
  • Parameter Count: 808,878,080
  • Quantization Precision: 8-bit
  • File Size: 833.69 MB
  • Format: GGML
  • License: mit
  • SHA256: 317eb69c11673c9de1e1f0d459b253999804ec71ac4c23c17ecf5fbe24e259a1
⬇️ Download whisper-large-v3-turbo-q8_0.bin
OpenAI Whisper Large V3whisper-large3

Description: The largest Whisper v3 model providing state-of-the-art transcription accuracy across varied audio.

Specifications:

  • Capabilities: Speech-to-Text
  • Architecture: whisper
  • Context Length: 1,500 tokens
  • Parameter Count: 1,543,490,560
  • Quantization Precision: 8-bit
  • File Size: 1.54 GB
  • Format: GGML
  • License: mit
  • SHA256: 37efc6b68f300ab717465685f7c3e175a66c11cf92bb3ab9912e86f4116c465e
⬇️ Download whisper-large-v3-q8_0.bin
OpenAI Whisper Mediumwhisper-medium

Description: A medium-sized Whisper model offering high-quality transcription for diverse audio scenarios.

Specifications:

  • Capabilities: Speech-to-Text
  • Architecture: whisper
  • Context Length: 1,500 tokens
  • Parameter Count: 763,857,920
  • Quantization Precision: 8-bit
  • File Size: 785.23 MB
  • Format: GGML
  • License: mit
  • SHA256: 42a1ffcbe4167d224232443396968db4d02d4e8e87e213d3ee2e03095dea6502
⬇️ Download whisper-medium-q8_0.bin
OpenAI Whisper Smallwhisper-small

Description: A small Whisper model providing improved transcription fidelity while remaining efficient.

Specifications:

  • Capabilities: Speech-to-Text
  • Architecture: whisper
  • Context Length: 1,500 tokens
  • Parameter Count: 241,734,912
  • Quantization Precision: 8-bit
  • File Size: 252.21 MB
  • Format: GGML
  • License: mit
  • SHA256: 49c8fb02b65e6049d5fa6c04f81f53b867b5ec9540406812c643f177317f779f
⬇️ Download whisper-small-q8_0.bin
OpenAI Whisper Tinywhisper-tiny

Description: The smallest Whisper variant offering fast, lightweight speech-to-text transcription.

Specifications:

  • Capabilities: Speech-to-Text
  • Architecture: whisper
  • Context Length: 1,500 tokens
  • Parameter Count: 37,760,640
  • Quantization Precision: 8-bit
  • File Size: 41.52 MB
  • Format: GGML
  • License: mit
  • SHA256: c2085835d3f50733e2ff6e4b41ae8a2b8d8110461e18821b09a15c40c42d1cca
⬇️ Download whisper-tiny-q8_0.bin