Which AI models can I run on an RTX 3060 12 GB?

413 of the 829 LLMs analyzed on Model Anatomy fit on an RTX 3060 12 GB (10.8 GB usable) with at least 4K tokens of context. Each row shows the best file that fits and how much context the remaining memory holds.

Hardware

How many

Context needed

Context cache

Sort by

Show models that don't fit

Model	Params	File that fits	Context that fits
Qwen/Qwen3-0.6BDense	752M596M active	As published1.50 GB of 10.8 GB available	Full 40K
Qwen/Qwen3-VL-8B-InstructDense + vision	8.77B7.57B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 39K tokensmax 256K
Qwen/Qwen3-8BDense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 40K
unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUFMoE	30.5B3.04B active	UD-IQ2_XXS10.3 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 5K tokensmax 256K
Qwen/Qwen2.5-7B-InstructDense	7.62B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
Qwen/Qwen3.5-9BDense + vision	9.65B8.18B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 159K tokensmax 256K
Qwen/Qwen2.5-0.5B-InstructDense	494M	As published988 MB of 10.8 GB available	Full 32K
Qwen/Qwen2.5-1.5B-InstructDense	1.54B	As published3.09 GB of 10.8 GB available	Full 32K
Qwen/Qwen3-4BDense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 40K
Qwen/Qwen3.5-4BDense + vision	4.66B	As published9.32 GB of 10.8 GB available	≈ 43K tokensmax 256K
Qwen/Qwen2.5-VL-7B-InstructDense + vision	8.29B7.07B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 104K tokensmax 125K
ornith-ai/Ornith-1.5-9B-GGUFDense	9.20B7.94B active	Q8_09.79 GB of 10.8 GB available	≈ 29K tokensmax 256K
Qwen/Qwen2.5-3B-InstructDense	3.09B	As published6.17 GB of 10.8 GB available	Full 32K
Qwen/Qwen3.5-2BDense + vision	2.27B1.94B active	As published4.55 GB of 10.8 GB available	Full 256K
Qwen/Qwen3-4B-Instruct-2507Dense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 256K
Qwen/Qwen3-1.7BDense	2.03B1.72B active	As published4.06 GB of 10.8 GB available	Full 40K
Qwen/Qwen3-VL-4B-InstructDense + vision	4.44B	As published8.88 GB of 10.8 GB available	≈ 13K tokensmax 256K
RadixArk/Kimi-K3-DSpark	2.25B	As published4.50 GB of 10.8 GB available	≈ 300K tokensmax 1M
Qwen/Qwen2.5-7B-Instruct-AWQDense	7.62B	As published5.57 GB of 10.8 GB available	Full 32K
nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16	3.97B	As published7.95 GB of 10.8 GB available	Not knowncontext figures missing
ornith-ai/Ornith-1.0-9B-GGUFDense	8.95B7.94B active	Q8_09.53 GB of 10.8 GB available	≈ 36K tokensmax 256K
Qwen/Qwen3-VL-2B-InstructDense + vision	2.13B1.72B active	As published4.26 GB of 10.8 GB available	≈ 56K tokensmax 256K
datalab-to/chandra-ocr-2Dense + vision	5.30B4.33B active	As published10.6 GB of 10.8 GB available	≈ 5K tokensmax 256K
Qwen/Qwen2.5-Coder-7B-InstructDense	7.62B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
Qwen/Qwen3-14B-AWQDense	14.8B	As published9.98 GB of 10.8 GB available	≈ 5K tokensmax 40K
Qwen/Qwen3-VL-8B-Instruct-FP8Dense + vision	8.77B7.57B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 39K tokensmax 256K
Qwen/Qwen2.5-VL-3B-InstructDense + vision	3.75B3.09B active	As published7.51 GB of 10.8 GB available	≈ 87K tokensmax 125K
Qwen/Qwen3.5-0.8BDense + vision	873M773M active	As published1.75 GB of 10.8 GB available	Full 256K
Qwen/Qwen2.5-14B-InstructDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 12K tokensmax 32K
deepseek-ai/DeepSeek-OCRMoE + vision	3.34B672M active	As published6.67 GB of 10.8 GB available	Full 8K
JonathanColetti/Qwen3.8-27B-Uncensored-GGUFDense	27.3B	Qwen3.8-27B-Uncensored-noMTP-IQ2_M10.2 GB of 10.8 GB available	≈ 7K tokensmax 256K
HauhauCS/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUFDense	27.3B	IQ2_M10.3 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 5K tokensmax 256K
huihui-ai/Huihui-Qwen3.8-27B-abliterated-GGUFDense	27.3B	Huihui-Qwen3.8-27B-abliterated-Swift-1.5-GSQ-RCO-IQ3_XXS10.1 GB of 10.8 GB available	≈ 8K tokensmax 256K
HuggingFaceTB/SmolLM2-135MDense	135M	As published269 MB of 10.8 GB available	Full 8K
DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP-GGUFDense	9.20B7.94B active	Qwen3.5-9B-The-Defiant-Fable-Uncnr-Heretic-NEO-MAX-Q8_010.5 GB of 10.8 GB available	≈ 8K tokensmax 256K
Qwen/Qwen2.5-14B-Instruct-AWQDense	14.8B	As published9.98 GB of 10.8 GB available	≈ 4K tokensmax 32K
Qwen/Qwen2-VL-2B-InstructDense + vision	2.21B1.54B active	As published4.42 GB of 10.8 GB available	Full 32K
vikhyatk/moondream2Dense + vision	1.93B1.37B active	As published3.85 GB of 10.8 GB available	Not knowncontext figures missing
Qwen/Qwen2.5-VL-7B-Instruct-AWQDense + vision	8.29B7.07B active	As published6.92 GB of 10.8 GB available	≈ 66K tokensmax 125K
HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive	7.52B	Q8_K8.13 GB of 10.8 GB available	≈ 88K tokensmax 128K
Qwen/Qwen3-14BDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 15K tokensmax 40K
Qwen/Qwen2.5-Coder-14B-InstructDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 12K tokensmax 32K
Qwen/Qwen3-8B-AWQDense	8.19B	As published6.10 GB of 10.8 GB available	≈ 31K tokensmax 40K
mistralai/Mistral-7B-Instruct-v0.2Dense	7.24B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
0bserverx/Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUFDense	26.9B	RVN-Q2_K_S-multilingual10.2 GB of 10.8 GB available	≈ 6K tokensmax 256K
prism-ml/Ternary-Bonsai-2-27B-ggufDense	26.9B	Ternary-Bonsai-2-27B-Q2_07.63 GB of 10.8 GB available	≈ 45K tokensmax 256K
llava-hf/llava-1.5-7b-hfDense + vision	7.06B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Not knowncontext figures missing
Qwen/Qwen2-VL-7B-Instruct-AWQDense + vision	8.29B7.07B active	As published6.92 GB of 10.8 GB available	Full 32K
Qwen/Qwen2.5-0.5BDense	494M	As published988 MB of 10.8 GB available	Full 32K
Qwen/Qwen3-4B-BaseDense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 32K
unsloth/Qwen3.5-9B-GGUFDense	8.95B7.94B active	Q8_09.53 GB of 10.8 GB available	≈ 36K tokensmax 256K
Qwen/Qwen3-1.7B-BaseDense	1.72B	As published3.44 GB of 10.8 GB available	Full 32K
HuggingFaceTB/SmolLM2-135M-InstructDense	135M	As published269 MB of 10.8 GB available	Full 8K
peculiar-ragdoll/Dirk-Qwen3.8-27B-GGUFDense	27.3B	UD-Q2_K_XL9.83 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 12K tokensmax 256K
apple/OpenELM-1_1B-Instruct	1.08B	As published2.16 GB of 10.8 GB available	Not knowncontext figures missing
sahilchachra/Unlimited-OCR-AWQMoE + vision	3.34B672M active	As published2.82 GB of 10.8 GB available	Full 32K
ibm-research/PowerMoE-3b	3.37B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 4K
unsloth/Qwen3.6-35B-A3B-GGUFMoE	34.7B2.95B active	UD-IQ1_M10.0 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 33K tokensmax 256K
HuggingFaceTB/SmolVLM2-500M-Video-InstructDense + vision	507M374M active	As published2.03 GB of 10.8 GB available	Full 8K
datalab-to/surya-ocr-2Dense + vision	686M586M active	As published1.37 GB of 10.8 GB available	Full 256K
LiquidAI/LFM2.5-2.6B-GGUF	2.70B	BF165.40 GB of 10.8 GB available	Full 128K
Qwen/Qwen2.5-Coder-14B-Instruct-AWQDense	14.8B	As published9.98 GB of 10.8 GB available	≈ 4K tokensmax 32K
ornith-ai/Ornith-1.0-9BDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
unsloth/Qwen3.6-27B-MTP-GGUFDense	27.3B	UD-IQ2_XXS9.57 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 16K tokensmax 256K
ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUFDense	27.3B	Qwen3.8-27B-GSQ-RCO-IQ3_XXS10.1 GB of 10.8 GB available	≈ 8K tokensmax 256K
baidu/Unlimited-OCRMoE + vision	3.34B672M active	As published6.67 GB of 10.8 GB available	Full 32K
datalab-to/surya-ocr-2-ggufDense	632M565M active	surya-21.27 GB of 10.8 GB available	Full 256K
unsloth/Qwen3.6-27B-GGUFDense	26.9B	UD-IQ2_XXS9.39 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 19K tokensmax 256K
openbmb/MiniCPM5-2BDense	2.52B2.25B active	As published5.03 GB of 10.8 GB available	Full 128K
mlx-community/Llama-3.1-8B-Instruct-4bitDense	8.03B	As published4.52 GB of 10.8 GB available	≈ 47K tokensmax 128K
DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUFDense	27.3B	Qwen3.8-27B-Turbo-Brilliance-50X-RI-GSQ-RCO-IQ3_XXS10.1 GB of 10.8 GB available	≈ 8K tokensmax 256K
legraphista/glm-4-9b-chat-IMat-GGUFDense	9.40B	Q8_09.99 GB of 10.8 GB available	≈ 19K tokensmax 128K
QuantTrio/Qwen3.5-9B-AWQDense + vision	9.65B8.18B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 159K tokensmax 256K
bartowski/endless-frontier_BigBang-v1-GGUFMoE	35.5B2.95B active	IQ2_XXS10.3 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 23K tokensmax 256K
Qwen/Qwen3-4B-Instruct-2507-FP8Dense	4.41B	As published5.19 GB of 10.8 GB available	≈ 37K tokensmax 256K
Qwen/Qwen2.5-Coder-7B-Instruct-AWQDense	7.62B	As published5.57 GB of 10.8 GB available	Full 32K
deepseek-ai/DeepSeek-Coder-V2-Lite-InstructMoE	15.7B2.45B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 62K tokensmax 160K
unsloth/Qwen3.5-4B-GGUFDense	4.21B	BF168.42 GB of 10.8 GB available	≈ 69K tokensmax 256K
CMSManhattan/JiRackUltra_1bDense	1.78B1.54B active	As published3.55 GB of 10.8 GB available	Full 128K
deepseek-ai/DeepSeek-OCR-2MoE	3.39B850M active	As published6.78 GB of 10.8 GB available	Full 8K
Qwen/Qwen2.5-1.5BDense	1.54B	As published3.09 GB of 10.8 GB available	Full 128K
bartowski/XYZAILab_XYZ-Aquila-mini-GGUFMoE	34.7B2.95B active	IQ2_XXS9.78 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 46K tokensmax 256K
agentionai/Qwen3.8-27B-AP-GGUFDense	27.3B	IQ2_S9.61 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 15K tokensmax 256K
deepseek-ai/DeepSeek-R1-0528-Qwen3-8BDense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 41K tokensmax 128K
peculiar-ragdoll/Sharp-Spark-X2.5-4B-GGUFDense	4.11B	Q6_K_XL3.61 GB of 10.8 GB available	≈ 189K tokensmax 1M
raxcore-dev/Rax-4.5Dense + vision	2.27B1.94B active	As published4.55 GB of 10.8 GB available	Full 256K
unsloth/gemma-4-12b-it-GGUFDense	11.9B	UD-Q5_K_XL8.61 GB of 10.8 GB available	≈ 11K tokensmax 256K
Qwen/Qwen2-VL-7B-InstructDense + vision	8.29B7.07B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
DavidAU/LFM2.5-24B-A2B-Turbo-Brilliance-50X-NEO-MAX-Imatrix-GGUFMoE	24.0B2.45B active	IQ3_M10.5 GB of 10.8 GB available	≈ 13K tokensmax 128K
prism-ml/Ternary-Bonsai-2-9B-gguf	1.47B	Ternary-Bonsai-2-9B-PQ2_0-MTP-Q8_02.68 GB of 10.8 GB available	Full 256K
dots-studio/dots.ocrDense + vision	3.04B1.54B active	As published6.08 GB of 10.8 GB available	Full 128K
allenai/OLMo-2-0425-1BDense	1.48B1.28B active	As published5.94 GB of 10.8 GB available	Full 4K
Qwen/Qwen3-0.6B-BaseDense	596M	As published1.19 GB of 10.8 GB available	Full 32K
empero-ai/Qwen3.8-4B-Distill-GGUFDense	4.33B	BF168.67 GB of 10.8 GB available	≈ 62K tokensmax 256K
Qwen/Qwen3-8B-FP8Dense	8.19B	As published9.44 GB of 10.8 GB available	≈ 9K tokensmax 40K
empero-ai/Qwen3.8-2B-Distill-GGUFDense	1.94B	BF163.90 GB of 10.8 GB available	Full 256K
FINAL-Bench/POCKET-35B-GGUFMoE	34.7B2.95B active	IQ1_M8.24 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 119K tokensmax 256K
Infomaniak-AI/vllm-translategemma-4b-itDense + vision	4.30B	As published8.60 GB of 10.8 GB available	≈ 99K tokensmax 128K
ConwayResearch/Underdog-Saluki-27B-1.0Dense	26.9B	Underdog-Saluki-27B-1.0-IQ2-mix7.90 GB of 10.8 GB available	≈ 41K tokensmax 256K
Qwen/Qwen2-0.5BDense	494M	As published988 MB of 10.8 GB available	Full 128K
LGAI-EXAONE/EXAONE-3.5-7.8B-Instruct-AWQDense	7.82B	As published5.30 GB of 10.8 GB available	Full 32K
deepseek-ai/deepseek-coder-7b-instruct-v1.5Dense	6.91B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 4K
Qwen/Qwen2.5-Coder-7BDense	7.62B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUFDense	11.9B	Q4_K_M7.38 GB of 10.8 GB available	≈ 20K tokensmax 256K
OpenGVLab/InternVL2-2B	2.21B	As published4.41 GB of 10.8 GB available	Full 32K
Qwen/Qwen1.5-MoE-A2.7BMoE	14.3B2.38B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 8K
empero-ai/Qwen3.8-9B-Distill-GGUFDense	9.20B7.94B active	Q8_09.79 GB of 10.8 GB available	≈ 29K tokensmax 256K
z-lab/Qwen3.8-27B-DFlash2-GGUF	1.92B	BF163.86 GB of 10.8 GB available	Full 256K
Salesforce/blip2-opt-2.7b	3.74B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 26K tokens
microsoft/Phi-3.5-vision-instructDense + vision	4.15B3.72B active	As published8.29 GB of 10.8 GB available	≈ 6K tokensmax 128K
jegreene/Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO-GGUFDense	27.3B	IQ2_S9.61 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 15K tokensmax 256K
Qwen/Qwen2.5-7BDense	7.62B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 111K tokensmax 128K
boboliu/bge-reranker-v2.5-gemma2-lightweight-gptqDense	9.24B	As published6.17 GB of 10.8 GB available	Full 8K
OpenGVLab/InternVL2-1BDense + vision	938M498M active	As published1.88 GB of 10.8 GB available	Full 32K
CMSManhattan/JiRackUltra_14bDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 12K tokensmax 128K
Orion-zhen/Qwen2.5-Coder-7B-Instruct-AWQDense	7.62B	As published5.70 GB of 10.8 GB available	Full 32K
microsoft/phi-4Dense	14.7B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 12K tokensmax 16K
openbmb/MiniCPM5-2B-DSpark	324M	As published648 MB of 10.8 GB available	Full 128K
tencent/HunyuanOCRDense + vision	1.12B950M active	As published2.24 GB of 10.8 GB available	≈ 85K tokensmax 128K
prism-ml/Ternary-Bonsai-27B-gguf	3.65B	Ternary-Bonsai-27B-Q2_g647.59 GB of 10.8 GB available	Full 4K
Qwen/Qwen2.5-VL-3B-Instruct-AWQDense + vision	3.75B3.09B active	As published3.40 GB of 10.8 GB available	Full 125K
OsaurusAI/Diplodocus-preview-GGUFDense	9.00B7.97B active	Q8_09.57 GB of 10.8 GB available	≈ 8K tokensmax 512K
Qwen/Qwen2.5-3B-Instruct-AWQDense	3.40B	As published2.69 GB of 10.8 GB available	Full 32K
lightonai/LightOnOCR-2-1BDense + vision	1.01B596M active	As published2.01 GB of 10.8 GB available	Full 16K
Qwen/Qwen2.5-7B-Instruct-GGUFDense	7.62B	Q8_08.10 GB of 10.8 GB available	≈ 46K tokensmax 128K
Qwen/Qwen2-1.5B-InstructDense	1.54B	As published3.09 GB of 10.8 GB available	Full 32K
unsloth/gemma-4-E2B-it-GGUF	4.65B	BF169.31 GB of 10.8 GB available	≈ 99K tokensmax 128K
hugging-quants/gemma-2-9b-it-AWQ-INT4Dense	9.24B	As published6.16 GB of 10.8 GB available	Full 8K
HuggingFaceTB/SmolLM3-3BDense	3.08B	As published6.15 GB of 10.8 GB available	≈ 62K tokensmax 64K
unsloth/Llama-3.2-1B-InstructDense	1.24B	As published2.47 GB of 10.8 GB available	Full 128K
bartowski/Qwen2.5-32B-Instruct-GGUFDense	32.8B	IQ2_XXS9.03 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 7K tokensmax 32K
dealignai/Ornith-1.5-9B-UNCENSORED-GGUFDense	8.95B7.94B active	Q8_09.53 GB of 10.8 GB available	≈ 36K tokensmax 256K
lmstudio-community/Qwen3.5-9B-MLX-4bitDense + vision	9.41B7.94B active	As published5.95 GB of 10.8 GB available	≈ 143K tokensmax 256K
cyankiwi/MiniCPM-SALA-AWQ-8bitDense	9.48B	As published10.6 GB of 10.8 GB available	≈ 5K tokensmax 512K
unsloth/gemma-4-26B-A4B-it-GGUFMoE	25.2B3.82B active	UD-IQ2_M10.0 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 4K tokensmax 256K
ibm-granite/granite-4.1-3bDense	3.40B	As published6.81 GB of 10.8 GB available	≈ 48K tokensmax 128K
unsloth/gemma-4-E4B-it-GGUF	7.52B	UD-Q8_K_XL8.71 GB of 10.8 GB available	≈ 69K tokensmax 128K
PrimeIntellect/Qwen3-0.6BDense	752M596M active	As published1.50 GB of 10.8 GB available	Full 40K
Qwen/Qwen3-14B-BaseDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 15K tokensmax 32K
lmstudio-community/Qwen3.5-9B-MLX-8bitDense + vision	9.41B7.94B active	As published10.4 GB of 10.8 GB available	≈ 10K tokensmax 256K
HauhauCS/Gemma4-12B-QAT-Uncensored-HauhauCS-BalancedDense	11.9B	Q4_K_M7.38 GB of 10.8 GB available	≈ 20K tokensmax 256K
empero-ai/Qwythos-9B-Claude-Mythos-5-1M-GGUFDense	9.20B7.94B active	Q8_09.79 GB of 10.8 GB available	≈ 29K tokensmax 1M
ornith-ai/Ornith-1.5-9BDense + vision	9.65B8.18B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 159K tokensmax 256K
Qwen/Qwen3.5-4B-BaseDense + vision	4.66B	As published9.32 GB of 10.8 GB available	≈ 43K tokensmax 256K
XHToken/Spark-X2.6-4B-GGUFDense	4.45B	Spark-X2.6-4B8.23 GB of 10.8 GB available	≈ 67K tokensmax 1M
Qwen/Qwen3-8B-BaseDense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5BDense	1.78B1.54B active	As published3.55 GB of 10.8 GB available	Full 128K
Qwen/Qwen3-4B-GGUFDense	4.02B	Q8_04.28 GB of 10.8 GB available	Full 40K
Qwen/Qwen3-4B-Thinking-2507Dense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 256K
Qwen/Qwen3-8B-GGUFDense	8.19B	Q8_08.71 GB of 10.8 GB available	≈ 14K tokensmax 40K
prism-ml/Bonsai-27B-gguf	3.65B	BF167.29 GB of 10.8 GB available	Full 4K
Qwen/Qwen3.5-9B-BaseDense + vision	9.65B8.18B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 159K tokensmax 256K
Qwen/Qwen3-VL-4B-Instruct-FP8Dense + vision	4.83B4.02B active	As published6.02 GB of 10.8 GB available	≈ 32K tokensmax 256K
mistralai/Mistral-7B-v0.1Dense	7.24B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
Qwen/Qwen3.5-2B-BaseDense + vision	2.27B1.94B active	As published4.55 GB of 10.8 GB available	Full 256K
deepseek-ai/DeepSeek-R1-Distill-Qwen-14BDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 12K tokensmax 128K
Qwen/Qwen2.5-3BDense	3.09B	As published6.17 GB of 10.8 GB available	Full 32K
Qwen/Qwen2.5-Coder-7B-Instruct-GGUFDense	7.62B	Q8_08.10 GB of 10.8 GB available	≈ 46K tokensmax 128K
ukisai/Swift-Qwen3.8-27B-GGUFDense	27.3B	IQ2_S9.90 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 11K tokensmax 256K
deepseek-ai/DeepSeek-R1-Distill-Qwen-7BDense	7.62B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 111K tokensmax 128K
bartowski/Qwen3.8-27B-GGUFDense	27.3B	IQ2_S9.68 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 14K tokensmax 256K
Qwen/Qwen3-0.6B-FP8Dense	752M596M active	As published1.06 GB of 10.8 GB available	Full 40K
Qwen/Qwen3.5-0.8B-BaseDense + vision	873M773M active	As published1.75 GB of 10.8 GB available	Full 256K
Qwen/Qwen3-4B-AWQDense	4.02B	As published2.67 GB of 10.8 GB available	Full 40K
XHToken/Spark-X2.5-4B-GGUFDense	4.11B	Spark-X2.5-4B8.23 GB of 10.8 GB available	≈ 67K tokensmax 1M
yuxinlu1/gemma-4-12B-coder-fable5-composer2.5-v1-GGUFDense	11.9B	Q4_K_M7.38 GB of 10.8 GB available	≈ 20K tokensmax 256K
Maximilian228/Qwen3.8-Flash-Next-NVFP4-GPTQ-StrataDense	636M	qwen-nvfp4-gptq-dense5.99 GB of 10.8 GB available	Not knowncontext figures missing
Qwen/Qwen3-0.6B-GGUFDense	596M	Q8_0639 MB of 10.8 GB available	Full 40K
Qwen/Qwen3-14B-GGUFDense	14.8B	Q4_K_M9.00 GB of 10.8 GB available	≈ 11K tokensmax 40K
deepseek-ai/DeepSeek-R1-Distill-Llama-8BDense	8.03B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 47K tokensmax 128K
moonshotai/Kimi-VL-A3B-InstructMoE + vision	16.4B2.58B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 49K tokensmax 128K
LiquidAI/LFM3-1.7B-GGUF	1.73B	BF163.47 GB of 10.8 GB available	Full 256K
openbmb/MiniCPM-V-4.6Dense + vision	1.30B778M active	As published2.60 GB of 10.8 GB available	Full 256K
DavidAU/Qwen3.8-27B-TWIN-TURBO-Fable-Cold-Fusion-709-L-Uncensored-NM-DAU-NEO-MTP-GGUFDense	27.3B	Qwen3.8-27B-TTURBO-Fable-C-Fusion-709-L-Uncen-NM-DAU-NEO-MTP-IQ2_M10.2 GB of 10.8 GB available	≈ 6K tokensmax 256K
JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUFMoE	12.1B2.21B active	Mellum2.1-12B-A2.5B-Thinking-MTP-Q4_K_M8.33 GB of 10.8 GB available	Full 128K
openbmb/MiniCPM5-2B-GGUFDense	2.52B2.25B active	F165.04 GB of 10.8 GB available	Full 128K
Jackrong/DeepSeek-V4-Pro-Qwen3.5-9B-MTP-GGUFDense	9.20B7.94B active	Q8_09.79 GB of 10.8 GB available	≈ 29K tokensmax 256K
Qwen/Qwen3-VL-8B-ThinkingDense + vision	8.77B7.57B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 39K tokensmax 256K
ReliquaryForge/qwen3-4b-base-dapo-v4Dense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 32K
baidu/Unlimited-OCR-2MoE + vision	3.34B680M active	As published6.69 GB of 10.8 GB available	≈ 65K tokensmax 128K
Qwen/Qwen3Guard-Gen-0.6BDense	752M596M active	As published1.50 GB of 10.8 GB available	Full 32K
Qwen/Qwen3-VL-8B-Instruct-GGUFDense	8.19B	Q8_08.71 GB of 10.8 GB available	≈ 14K tokensmax 256K
Qwen/Qwen3-VL-4B-Instruct-GGUFDense	4.02B	F168.05 GB of 10.8 GB available	≈ 18K tokensmax 256K
Qwen/Qwen3-14B-FP8Dense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 15K tokensmax 40K
Qwen/Qwen3-VL-2B-Instruct-GGUFDense	1.72B	F163.45 GB of 10.8 GB available	≈ 63K tokensmax 256K
LiquidAI/LFM2.5-2.6B	2.70B	As published5.39 GB of 10.8 GB available	Full 128K
cyankiwi/Qwen3-VL-8B-Instruct-AWQ-4bitDense + vision	8.77B7.57B active	As published7.55 GB of 10.8 GB available	≈ 22K tokensmax 256K
Qwen/Qwen3-VL-2B-ThinkingDense + vision	2.13B1.72B active	As published4.26 GB of 10.8 GB available	≈ 56K tokensmax 256K
Qwen/Qwen3-VL-2B-Instruct-FP8Dense + vision	2.44B1.72B active	As published3.47 GB of 10.8 GB available	≈ 62K tokensmax 256K
Qwen/Qwen3-4B-FP8Dense	4.41B	As published5.19 GB of 10.8 GB available	≈ 37K tokensmax 40K
nvidia/LocateAnything-3BDense + vision	3.83B3.10B active	As published7.66 GB of 10.8 GB available	Full 32K
ukisai/Swift-1.5-Qwen3.8-27B-GGUFDense	27.3B	IQ2_S9.68 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 14K tokensmax 256K
Qwen/Qwen2.5-14BDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 12K tokensmax 128K
LiquidAI/LFM3-1.7B	1.73B	As published3.46 GB of 10.8 GB available	Full 256K
Qwen/Qwen3-1.7B-GGUFDense	1.72B	Q8_01.83 GB of 10.8 GB available	Full 40K
LiquidAI/LFM3-VL-2B-GGUF	1.73B	BF163.47 GB of 10.8 GB available	Full 125K
jialinyyzz/humanizerDense + vision	12.0B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 56K tokensmax 256K
Qwen/Qwen3Guard-Gen-4BDense	4.41B	As published8.82 GB of 10.8 GB available	≈ 13K tokensmax 32K
Qwen/Qwen3-VL-4B-ThinkingDense + vision	4.44B	As published8.88 GB of 10.8 GB available	≈ 13K tokensmax 256K
openbmb/MiniCPM-V-4.7-1B-ggufDense	1.01B752M active	F162.02 GB of 10.8 GB available	Full 256K
IFM/K2-Horizon-7B-UnoDense	349M	As published1.40 GB of 10.8 GB available	Not knowncontext figures missing
unsloth/LFM2.5-24B-A2B-GGUFMoE	24.0B2.45B active	UD-IQ3_XXS8.96 GB of 10.8 GB available	≈ 88K tokensmax 128K
huihui-ai/Huihui-DeepSeek-V4.1-Flash-1017-abliterated-GGUFDense	755M	Huihui-DeepSeek-V4.1-Flash-1017-patch-Q2-13.66 GB of 10.8 GB available	Not knowncontext figures missing
XingChen-AGI/TeleOCRDense + vision	1.42B596M active	As published2.83 GB of 10.8 GB available	≈ 68K tokensmax 125K
RavichandranJ/Dolphin3-Cyber-8B-GGUFDense	8.03B	Q8_08.54 GB of 10.8 GB available	≈ 17K tokensmax 128K
DavidAU/LFM2.5-2.6B-Qwen3.8-Turbo-Brilliance-Power-X12-NEO-MAX-GGUF	2.70B	BF165.40 GB of 10.8 GB available	Full 128K
Qwen/Qwen3Guard-Gen-8BDense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-GGUFDense	3.98B	Q5_K_M3.67 GB of 10.8 GB available	Not knowncontext figures missing
ggml-org/Clef-Flash-GGUFDense	9.08B8.06B active	Q8_09.66 GB of 10.8 GB available	≈ 33K tokensmax 256K
PaddlePaddle/PaddleOCR-VL-1.6Dense + vision	959M387M active	As published1.92 GB of 10.8 GB available	Full 128K
Cloudflare/clef-flashDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
openbmb/MiniCPM-V-4.7-1BDense + vision	1.55B778M active	As published3.11 GB of 10.8 GB available	Full 256K
EldanRing/Winnow-12BDense	11.9B	Winnow-12B-NVFP48.16 GB of 10.8 GB available	≈ 15K tokensmax 256K
moonshotai/Moonlight-16B-A3B-InstructMoE	16.0B2.58B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 8K
StarDoc-AI/TeleOCRDense + vision	1.42B596M active	As published2.83 GB of 10.8 GB available	≈ 68K tokensmax 125K
XHToken/Spark-X2.5-4B	4.11B	As published8.22 GB of 10.8 GB available	≈ 67K tokensmax 1M
DavidAU/Qwen3.8-27B-Ultimate-Heretic-Uncensored-Turbo-Brilliance-50x-AGI-NEO-GGUFDense	27.3B	IQ2_M10.2 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 6K tokensmax 256K
IFM/K2-Horizon-7BDense	9.00B7.97B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 38K tokensmax 512K
moonshotai/Kimi-VL-A3B-Thinking-2506MoE + vision	16.4B2.58B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 49K tokensmax 128K
byteshape/Qwen3.8-27B-GGUFDense	27.3B	IQ3_XS10.4 GB of 10.8 GB available	≈ 4K tokensmax 256K
Qwen/Qwen3-VL-8B-Thinking-FP8Dense + vision	8.77B7.57B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 39K tokensmax 256K
TokenRhythm/NeoHorse-1-4BDense	4.21B	As published8.41 GB of 10.8 GB available	≈ 70K tokensmax 256K
bartowski/ukisai_Swift-Qwen3.8-27b-GGUFDense	27.3B	IQ2_S9.68 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 14K tokensmax 256K
TokenRhythm/NeoHorse-1-4B-GGUFDense	4.21B	BF168.42 GB of 10.8 GB available	≈ 69K tokensmax 256K
IFM/K2-Horizon-7B-GGUFDense	9.00B7.97B active	Q8_09.57 GB of 10.8 GB available	≈ 8K tokensmax 512K
webAI-Official/TwIL-8BDense	8.79B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 35K tokensmax 128K
Qwen/Qwen3-1.7B-FP8Dense	2.03B1.72B active	As published2.65 GB of 10.8 GB available	Full 40K
IFM/K2-Horizon-3.7BDense	5.06B4.42B active	As published10.1 GB of 10.8 GB available	≈ 5K tokensmax 512K
inclusionAI/Ling-3.0-tinyMoE	7.89B1.14B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Not knowncontext figures missing
XHToken/Spark-X2.6-4B	4.11B	As published8.22 GB of 10.8 GB available	≈ 67K tokensmax 1M
Qwen/Qwen3-VL-2B-Thinking-FP8Dense + vision	2.44B1.72B active	As published3.47 GB of 10.8 GB available	≈ 62K tokensmax 256K
bartowski/FrogNano-4B-2609-GGUFDense	4.33B	BF168.67 GB of 10.8 GB available	≈ 62K tokensmax 256K
deepgrove/maple-preview-GGUFMoE	20.2B1.18B active	maple-preview-TQ2_0-head-F166.35 GB of 10.8 GB available	Full 128K
LiquidAI/d1-2B-GGUF	1.73B	BF163.47 GB of 10.8 GB available	Full 125K
XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9BDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
openbmb/MiniCPM-OCR-1.0Dense + vision	1.30B778M active	As published2.60 GB of 10.8 GB available	Full 256K
paradigma-inc/limite-1b-violettoDense	1.04B	As published2.07 GB of 10.8 GB available	Full 128K
moonshotai/Kimi-VL-A3B-ThinkingMoE + vision	16.4B2.58B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 49K tokensmax 128K
xiaomi-research/Xiaomi-OCR-0Dense + vision	873M773M active	As published1.75 GB of 10.8 GB available	Full 256K
moonshotai/Moonlight-16B-A3BMoE	16.0B2.58B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 8K
0bserverx/Qwen3.8-27B-Heretic-GSQ-RCO-GGUFDense	27.3B	RVN-Qwen3.8-27B-Heretic-GSQ-RCO-IQ3_XXS10.1 GB of 10.8 GB available	≈ 8K tokensmax 256K
SC117/Spark-X2.6-4B-abliterated-FIT-GGUFDense	4.11B	BF168.23 GB of 10.8 GB available	≈ 67K tokensmax 1M
IFM/K2-Horizon-0.9BDense	1.08B	As published2.16 GB of 10.8 GB available	Full 128K
bartowski/Cloudflare_clef-flash-GGUFDense	9.08B8.06B active	Q8_09.68 GB of 10.8 GB available	≈ 32K tokensmax 256K
deepseek-ai/DeepSeek-Coder-V2-Lite-BaseMoE	15.7B2.45B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 62K tokensmax 160K
ccore/Qwen3.5-4B-GLB-3D-CoderDense + vision	4.54B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 244K tokensmax 256K
TokenRhythm/NeoHorse-1-9BDense	8.95B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 170K tokensmax 256K
LiquidAI/LFM3-VL-2B	1.92B	As published3.84 GB of 10.8 GB available	Full 125K
tencent/Hy-Memory-30B-A3B-GGUFMoE	30.1B3.20B active	Hy-Memory-30B-A3B-Q8_0-expsIQ2_XXS9.14 GB of 10.8 GB available	≈ 16K tokensmax 256K
Alissonerdx/Molmo2-8B-Minimax-H3Dense + vision	16.7B	As published9.29 GB of 10.8 GB available	≈ 10K tokensmax 36K
nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16	3.98B	As published7.95 GB of 10.8 GB available	≈ 14K tokensmax 256K
openbmb/MiniCPM-OCR-1.0-ggufDense	752M	F161.52 GB of 10.8 GB available	Full 256K
nvidia/Nemotron-3.5-Parse	631M	As published2.52 GB of 10.8 GB available	Full 16K
huihui-ai/Huihui-GLM-5.3-abliterated-GGUFDense	9.14B	Huihui-GLM-5.3-abliterated-GGUF-patch-Q2-16.45 GB of 10.8 GB available	Not knowncontext figures missing
abenzerps/Nanbeige4.3-33B-GSQ-RCO-GGUFDense	33.1B	IQ2_XXS9.90 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 4K tokensmax 512K
Kujira/Underdog-Saluki-27B-1.0-MTP-Abliterated-GGUFDense	27.3B	Underdog-Saluki-27B-1.0-IQ2-mix-MTP-abliterated8.35 GB of 10.8 GB available	≈ 34K tokensmax 256K
Qwen/Qwen3-4B-Thinking-2507-FP8Dense	4.41B	As published5.19 GB of 10.8 GB available	≈ 37K tokensmax 256K
peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUFDense	2.52B2.25B active	Q8_K_M2.93 GB of 10.8 GB available	Full 128K
bartowski/MiMo-V2.6-Distill-Qwen-9B-GGUFDense	8.95B7.94B active	Q8_09.55 GB of 10.8 GB available	≈ 36K tokensmax 256K
DavidAU/LFM2.5-8B-A1B-Qwen3.8-Turbo-Brilliance-Power-X12-NEO-MAX-GGUFMoE	8.47B1.69B active	Q8_09.26 GB of 10.8 GB available	≈ 123K tokensmax 125K
XHToken/Spark-X2.6-1.7B-GGUFDense	1.98B1.71B active	Spark-X2.6-1.7B3.42 GB of 10.8 GB available	≈ 501K tokensmax 1M
AtomicChat/LFM3-1.7B-GGUF	1.73B	Q8_01.85 GB of 10.8 GB available	Full 256K
Kujira/Underdog-Saluki-27B-1.0-MTP-GGUFDense	27.3B	Underdog-Saluki-27B-1.0-IQ2-mix-MTP8.35 GB of 10.8 GB available	≈ 34K tokensmax 256K
xtools-at/DeepSeek-V4-Pro-Qwen3.5-4B-UNCENSORED-hereticDense + vision	4.66B	As published9.32 GB of 10.8 GB available	≈ 43K tokensmax 256K
Edge0/Edge0-8B-A1B-previewMoE	7.95B1.20B active	As published4.56 GB of 10.8 GB available	Not knowncontext figures missing
LiquidAI/d1-2B	1.92B	As published3.84 GB of 10.8 GB available	Full 125K
ProCreations/Ternary-Bonsai-2-27B-MTP	425M	As published849 MB of 10.8 GB available	Not knowncontext figures missing
h2oai/h2o-lightning-4bDense + vision	4.66B	As published9.32 GB of 10.8 GB available	≈ 43K tokensmax 256K
Qwen/Qwen-Drive-1.0-4BDense + vision	4.54B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Not knowncontext figures missing
DeepHat/DeepHat-V1-7BDense	7.62B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
bartowski/NousResearch_Arete-27B-GGUFDense	26.9B	IQ2_M10.3 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 5K tokensmax 256K
saidutta69/Qwen3.5-4B-hereticDense + vision	4.54B	As published9.08 GB of 10.8 GB available	≈ 50K tokensmax 256K
LiquidAI/LFM3-1.7B-Base	1.73B	As published3.46 GB of 10.8 GB available	Full 256K
Qwen/Qwen3-VL-8B-Thinking-GGUFDense	8.19B	Q8_08.71 GB of 10.8 GB available	≈ 14K tokensmax 256K
ATH-MaaS/OvisOCR3Dense + vision	853M752M active	As published1.71 GB of 10.8 GB available	Full 256K
saidutta69/MiniCPM-V-4.7-1B-hereticDense + vision	1.55B778M active	As published3.11 GB of 10.8 GB available	Full 256K
Qwen/Qwen2-VL-2BDense + vision	2.21B1.54B active	As published4.42 GB of 10.8 GB available	Full 32K
bartowski/Altworld_Hemmingway-1-GGUFDense	27.3B	IQ2_S9.68 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 14K tokensmax 256K
SeerRay-Lab/Xiaomi-OCR-0Dense + vision	873M773M active	As published1.75 GB of 10.8 GB available	Full 256K
Qwen/Qwen2.5-Coder-14BDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 12K tokensmax 32K
openbmb/MiniCPM-SALADense	9.48B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 512K
tencent/WeVisOne-VL-2BDense + vision	2.44B1.72B active	As published4.88 GB of 10.8 GB available	≈ 50K tokensmax 256K
webAI-Official/TwIL-LM3-ProDense	3.66B	As published7.32 GB of 10.8 GB available	≈ 41K tokensmax 128K
Jackrong/Qwopus3-5-9B-Coder-Flash-GGUFDense	9.20B7.94B active	Q8_09.79 GB of 10.8 GB available	≈ 29K tokensmax 256K
TaichuAI/ZDTaichu5.0-9BDense + vision	9.79B8.13B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 156K tokensmax 256K
Qwen/Qwen3-VL-4B-Thinking-GGUFDense	4.02B	F168.05 GB of 10.8 GB available	≈ 18K tokensmax 256K
tencent/Youtu-Parsing-OmniDense + vision	5.33B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 165K tokensmax 1M
internlm/Intern-Decision-0.8BDense + vision	853M752M active	As published1.71 GB of 10.8 GB available	Full 256K
zgcagi/ZGCM-1-7BDense	7.39B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 256K
basecompute/Qwen3.8-27B-BaseQuant-V1Dense	26.9B	Qwen3.8-27B-BaseQuant-V1-XS8.36 GB of 10.8 GB available	≈ 34K tokensmax 256K
Hikari07jp/MiMo-V2.6-Distill-Qwen-9B-AblitratedDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
jialinyyzz/humanizer-GGUFDense	11.9B	Q4_K_M7.63 GB of 10.8 GB available	≈ 19K tokensmax 256K
XHToken/Spark-X2.6-1.7B	1.71B	As published3.42 GB of 10.8 GB available	≈ 502K tokensmax 1M
empero-ai/Qwen3.8-2B-DistillDense + vision	2.27B1.94B active	As published4.55 GB of 10.8 GB available	Full 256K
bytkim/Qwen3.8-27B-pi-GGUFDense	26.9B	IQ2_M10.0 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 10K tokensmax 256K
HuggingFaceBio/Carbon-XDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 15K tokensmax 40K
openbmb/MiniCPM-V-4.7-35B-A3B-DSpark	389M	As published778 MB of 10.8 GB available	Full 256K
tencent/WeVisOne-VL-4BDense + vision	4.83B4.02B active	As published9.66 GB of 10.8 GB available	≈ 8K tokensmax 256K
kaushikvira/DeepSeek-V4.1-Flash-1017-GSQ-agentic-strataDense + vision	14.7B	Q2_K5.66 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	Not knowncontext figures missing
OrionLLM/OxCoder-9BDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
mohit67890/imajev-4bDense	123M	As published490 MB of 10.8 GB available	Not knowncontext figures missing
taobao-mnn/MNN-Pocket-0.3BDense	362M303M active	As published723 MB of 10.8 GB available	Full 40K
AtomicChat/LFM3-1.7B-Uncensored-GGUF	1.73B	Q6_K1.59 GB of 10.8 GB available	Full 256K
ggml-org/MiMo-V2.6-Distill-Qwen-9B-GGUFDense	8.95B7.94B active	Q8_09.53 GB of 10.8 GB available	≈ 36K tokensmax 256K
openbmb/MiniCPM-OCR-JevDense + vision	2.21B1.88B active	As published4.43 GB of 10.8 GB available	Full 256K
XHToken/Spark-X2.6-4B-Base	4.11B	As published8.22 GB of 10.8 GB available	≈ 67K tokensmax 1M
aj9o9/Qwen3.8-27B-DSpine-b16-GGUF	2.03B	BF164.06 GB of 10.8 GB available	Full 256K
Qwen/Qwen3-14B-MLX-4bitDense	14.8B	As published7.85 GB of 10.8 GB available	≈ 18K tokensmax 40K
emperorofrome/GmcoderDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
XHToken/Spark-X2.6-4B-DSpark-GGUF	624M	Spark-X2.6-4B-DSpark1.25 GB of 10.8 GB available	Full 1M
DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTPDense + vision	9.65B8.18B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 159K tokensmax 256K
XHToken/Spark-X2.6-4B-FP8	4.11B	As published4.45 GB of 10.8 GB available	≈ 167K tokensmax 1M
Qwen/Qwen3-8B-MLX-4bitDense	8.19B	As published4.35 GB of 10.8 GB available	Full 40K
ceia-br/lumi-4b	3.97B	As published7.95 GB of 10.8 GB available	Not knowncontext figures missing
Qwen/Qwen3-VL-2B-Thinking-GGUFDense	1.72B	F163.45 GB of 10.8 GB available	≈ 63K tokensmax 256K
togethercomputer/Tev1-4B-experimentalDense + vision	4.66B	As published9.32 GB of 10.8 GB available	≈ 43K tokensmax 256K
divagr1925/Qwen3.5-4B-QwenputerDense	9.36B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
sensenova/SenseNova-Loop-9B-PreviewDense + vision	9.66B8.18B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 158K tokensmax 256K
tencent/Qwen3.8-27B-DSpine-b16	2.03B	As published4.05 GB of 10.8 GB available	Full 256K
jinaai/jina-ocr-v1MoE + vision	3.37B708M active	As published6.74 GB of 10.8 GB available	Full 32K
XHToken/Spark-X2.5-4B-Base	4.11B	As published8.22 GB of 10.8 GB available	≈ 67K tokensmax 1M
Qwen/Qwen3-4B-MLX-4bitDense	4.02B	As published2.14 GB of 10.8 GB available	≈ 57K tokensmax 64K
dealignai/Nanbeige4.3-33B-UNCENSORED-GGUFDense	33.1B	IQ2_XXS9.33 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 7K tokensmax 512K
xtools-at/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-hereticDense + vision	9.65B8.18B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 159K tokensmax 256K
Qwen/Qwen3-0.6B-MLX-4bitDense	596M	As published317 MB of 10.8 GB available	Full 38K
datalab-to/surya-ocr-2.5Dense + vision	689M589M active	As published1.38 GB of 10.8 GB available	Full 256K
Qwen/Qwen2-VL-7BDense + vision	8.29B7.07B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
Qwen/WebWorld-8BDense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 40K
Qwen/Qwen3-4B-SafeRLDense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 40K
XHToken/Spark-X2.6-1.7B-Base	1.71B	As published3.42 GB of 10.8 GB available	≈ 502K tokensmax 1M
Jackrong/DeepSeek-V4-Pro-Qwen3.5-9BDense + vision	9.65B8.18B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 159K tokensmax 256K
IFM/K2-Type-0.9BDense	1.08B	As published2.16 GB of 10.8 GB available	Full 128K
Octen/Octen-WebGR-0.6BDense	730M	As published1.46 GB of 10.8 GB available	Full 32K
lukey03/Qwen3.5-9B-abliteratedDense	8.95B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 170K tokensmax 256K
microsoft/Sentiment-Analysis-7BDense	7.24B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 32K
Octen/Octen-WebGR-4BDense	4.36B	As published8.72 GB of 10.8 GB available	≈ 14K tokensmax 32K
p-e-w/Oc-e-Non-v1Dense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 256K
GroundingPI/GroundAnything-VLMDense + vision	4.84B4.02B active	As published9.69 GB of 10.8 GB available	≈ 7K tokensmax 256K
Qwen/Qwen3-VL-4B-Thinking-FP8Dense + vision	4.83B4.02B active	As published6.02 GB of 10.8 GB available	≈ 32K tokensmax 256K
google/DiarizationLM-Gemma-4-E4B-v1Dense + vision	8.00B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 128K
allenai/AstaBrief_8BDense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 40K
Qwen/Qwen3-1.7B-MLX-4bitDense	1.72B	As published914 MB of 10.8 GB available	Full 64K
mlabonne/LFM2.5-230M-Chess	231M	As published461 MB of 10.8 GB available	Full 125K
Qwen/WebWorld-14BDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 15K tokensmax 40K
Qwen/Qwen3-0.6B-GPTQ-Int8Dense	596M	As published763 MB of 10.8 GB available	Full 40K
LiquidAI/d1-omni-600M-GGUF	381M	BF16764 MB of 10.8 GB available	Full 125K
abacusai/Smaug-HumanizerDense + vision	12.0B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 56K tokensmax 256K
Mapika/decider-2b-visionDense + vision	2.21B1.88B active	As published4.43 GB of 10.8 GB available	Full 256K
Cactus-Compute/mirage	54.1M	As published216 MB of 10.8 GB available	Not knowncontext figures missing
Qwen/Qwen3-1.7B-GPTQ-Int8Dense	1.72B	As published2.07 GB of 10.8 GB available	Full 40K
Qwen/Qwen3-8B-MLX-8bitDense	8.19B	As published8.45 GB of 10.8 GB available	≈ 16K tokensmax 40K
amazon/ALoDLM-8BDense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 40K
Qwen/Qwen3-14B-MLX-6bitDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 15K tokensmax 40K
LiquidAI/LFM2.5-VL-3B-DSpark	279M	As published559 MB of 10.8 GB available	Full 125K
scratchyone/anon-1bDense	1000M	As published2.60 GB of 10.8 GB available	Full 32K
maritaca-ai/S5-nanoDense + vision	4.66B	As published9.32 GB of 10.8 GB available	≈ 43K tokensmax 256K
ibm-granite/granite-switch-4.1-8b	8.78B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Not knowncontext figures missing
Qwen/Qwen3-14B-MLX-8bitDense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 15K tokensmax 40K
manuelgscom/dale-0.8b-GGUFDense	752M	Q4_K_M529 MB of 10.8 GB available	Full 256K
SeerRay-Lab/Xiaomi-Lens-0Dense + vision	4.62B	As published9.24 GB of 10.8 GB available	≈ 45K tokensmax 256K
Qwen/Qwen3-0.6B-MLX-8bitDense	596M	As published615 MB of 10.8 GB available	Full 40K
LastMachine/Qwen3-8B-ZeroZero-HereticDense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 40K
Qwen/Qwen3-1.7B-MLX-8bitDense	1.72B	As published1.77 GB of 10.8 GB available	Full 40K
169Pi/Alpie_seedDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
Qwen/Qwen3-8B-MLX-6bitDense	8.19B	As published6.40 GB of 10.8 GB available	≈ 29K tokensmax 40K
Qwen/Qwen3-4B-MLX-8bitDense	4.02B	As published4.15 GB of 10.8 GB available	Full 40K
nvidia/NV-Reason-CTDense + vision	5.32B4.21B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 231K tokensmax 256K
lightonai/LightOnOCR-3-1BDense + vision	1.01B596M active	As published2.01 GB of 10.8 GB available	Full 16K
Qwen/Qwen3-1.7B-MLX-bf16Dense	1.72B	As published3.44 GB of 10.8 GB available	Full 40K
deepseek-ai/ESFT-token-translation-liteMoE	865M81.1M active	As published1.73 GB of 10.8 GB available	Not knowncontext figures missing
deepseek-ai/ESFT-token-summary-liteMoE	1.16B109M active	As published2.32 GB of 10.8 GB available	Not knowncontext figures missing
Qwen/Qwen3-8B-MLX-bf16Dense	8.19B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 40K
Qwen/Qwen3-1.7B-MLX-6bitDense	1.72B	As published1.34 GB of 10.8 GB available	Full 40K
JonathanColetti/Qwen3.5-0.8B-Astra-WeightEditDense + vision	873M773M active	As published1.75 GB of 10.8 GB available	Full 256K
microsoft/AesCode-8BDense + vision	8.77B7.57B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 39K tokensmax 256K
Qwen/Qwen3-0.6B-MLX-6bitDense	596M	As published466 MB of 10.8 GB available	Full 40K
deepseek-ai/ESFT-token-law-liteMoE	1.48B139M active	As published2.96 GB of 10.8 GB available	Not knowncontext figures missing
deepseek-ai/ESFT-token-code-liteMoE	1.94B182M active	As published3.88 GB of 10.8 GB available	Not knowncontext figures missing
Qwen/Qwen3-4B-MLX-6bitDense	4.02B	As published3.14 GB of 10.8 GB available	Full 40K
deepseek-ai/ESFT-token-math-liteMoE	1.38B129M active	As published2.75 GB of 10.8 GB available	Not knowncontext figures missing
deepseek-ai/ESFT-vanilla-liteMoE	15.7B2.45B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 4K
deepseek-ai/ESFT-token-intent-liteMoE	1.60B150M active	As published3.20 GB of 10.8 GB available	Not knowncontext figures missing
deepseek-ai/ESFT-gate-law-liteMoE	1.65B155M active	As published3.30 GB of 10.8 GB available	Not knowncontext figures missing
deepseek-ai/ESFT-gate-code-liteMoE	2.34B220M active	As published4.69 GB of 10.8 GB available	Not knowncontext figures missing
Qwen/Qwen3-14B-MLX-bf16Dense	14.8B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 15K tokensmax 40K
deepseek-ai/ESFT-gate-translation-liteMoE	1.06B98.9M active	As published2.11 GB of 10.8 GB available	Not knowncontext figures missing
deepseek-ai/ESFT-gate-summary-liteMoE	1.74B163M active	As published3.48 GB of 10.8 GB available	Not knowncontext figures missing
deepseek-ai/ESFT-gate-math-liteMoE	1.58B148M active	As published3.17 GB of 10.8 GB available	Not knowncontext figures missing
deepseek-ai/ESFT-gate-intent-liteMoE	2.14B200M active	As published4.27 GB of 10.8 GB available	Not knowncontext figures missing
RikiyaT/Qwen3.5-0.8B-weight-space-programmingDense + vision	873M773M active	As published1.75 GB of 10.8 GB available	Full 256K
gugujiji/novel-humanize-4bDense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 40K
Qwen/Qwen3-0.6B-MLX-bf16Dense	596M	As published1.19 GB of 10.8 GB available	Full 40K
nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-BF16Dense + vision	8.00B	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 64K
Qwen/Qwen3-4B-MLX-bf16Dense	4.02B	As published8.04 GB of 10.8 GB available	≈ 18K tokensmax 40K
apple/LensVLM-9BDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
notnotsamuel/LFM2.5-350M-RLCD	354M	As published709 MB of 10.8 GB available	Full 125K
nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-4GB	7.52B	Q4_K_M6.18 GB of 10.8 GB available	Full 64K
lightonai/colgrep-agent-2BDense	2.52B2.25B active	As published5.03 GB of 10.8 GB available	Full 128K
JetBrains/Mellum2.1-12B-A2.5B-ThinkingMoE	12.1B2.21B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	Full 128K
lightonai/LightOnOCR-3-0.8BDense + vision	853M752M active	As published1.71 GB of 10.8 GB available	Full 256K
lightonai/LightOnOCR-3-4BDense + vision	4.54B	As published9.08 GB of 10.8 GB available	≈ 50K tokensmax 256K
AxiomicLabs/GPT-X3-150MDense	149M	As published598 MB of 10.8 GB available	Not knowncontext figures missing
ML-Intern-lab/Qwen3.5-0.8B-weight-space-programmingDense + vision	873M773M active	As published1.75 GB of 10.8 GB available	Full 256K
LiquidAI/d1-omni-600M	587M	As published2.35 GB of 10.8 GB available	Full 125K
LiquidAI/d1-3B-GGUF	2.70B	BF165.40 GB of 10.8 GB available	Full 32K
LiquidAI/d1-3B	3.12B	As published6.25 GB of 10.8 GB available	Full 32K
black-forest-labs/flux-3-image-klein-pupDense + vision	9.41B7.94B active	4-bit (estimate)no such file here: roughly what a 4-bit GGUF would take	≈ 163K tokensmax 256K
DavidAU/Qwen3.8-27B-Jumbo-Fusion-IQ-747-900-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUFDense	27.3B	IQ2_M10.2 GB of 10.8 GB availableunder 3 bits: noticeably lower quality	≈ 6K tokensmax 256K
unsloth/NVIDIA-Nemotron-3.5-Bolt-E4B-GGUF	7.52B	UD-Q8_K_XL8.19 GB of 10.8 GB available	≈ 86K tokensmax 128K
internlm/Intern-Decision-4BDense + vision	4.54B	As published9.08 GB of 10.8 GB available	≈ 50K tokensmax 256K
incoai/Qwen3.8-Flash-Next-DFlash2	603M	As published1.21 GB of 10.8 GB available	Full 256K
ukisai/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUFDense	27.3B	Swift-1.5-Qwen3.8-27B-GSQ-RCO-IQ3_XXS10.1 GB of 10.8 GB available	≈ 8K tokensmax 256K
microsoft/Mage-VL-CoworkDense + vision	4.74B4.02B active	As published9.48 GB of 10.8 GB available	≈ 9K tokensmax 256K

Assumes 90% of each card's memory is usable. Context memory is for one conversation. GGUF repos use their largest file that fits; other repos use the published weights, or a 4-bit estimate when only that would fit.