Models that run on 128 GB

1133 models with published weights that fit in 128 GB — a large Mac or a server card. Counted at four-bit quantisation, weights only, leaving the machine about a third of its memory and a gigabyte for context: room for roughly 136 billion parameters. A mixture of experts is counted in full, because it is held in full even though only a few experts compute. 1,133 in all, a hundred to a page; this is page 3 of 12.

A large Mac or a server card. This is where the biggest openly published models come within reach — and where the distinction between total and active parameters starts to decide everything, because a mixture of experts computes with a fraction of itself and must still be held whole.

• Qwen3 Omni 30B A3B InstructAlibaba30.0B≈19.5 GB at 4-bit2 also selling it hosted
• Qwen3 Omni 30B A3B ThinkingAlibaba30.0B≈19.5 GB at 4-bit2 also selling it hosted
• QwenLong-L1.5-30B-A3BTongyi-Zhiwen30.0B≈19.5 GB at 4-bit
• TildeOpen-30bTildeAI30.0B≈19.5 GB at 4-bit
• Wizard-Vicuna-30B-UncensoredQuixi AI30.0B≈19.5 GB at 4-bit
• WizardLM-30B-UncensoredQuixi AI30.0B≈19.5 GB at 4-bit
• granite-4.1-30bIBM Granite30.0B≈19.5 GB at 4-bit
• Muse Glimmer 30BMeta29.8B≈19.4 GB at 4-bit8 also selling it hosted
• stepvideo-t2vStepFun29.3B≈19.1 GB at 4-bit
• medgemma-27b-itGoogle28.8B≈18.7 GB at 4-bit
• translategemma-27b-itGoogle28.8B≈18.7 GB at 4-bit
• ERNIE 4.5 VL 28B A3BBaidu28.0B≈18.2 GB at 4-bit2 also selling it hosted
• ERNIE-4.5-VL-28B-A3B-ThinkingBaidu28.0B≈18.2 GB at 4-bit1 also selling it hosted
• Huihui-Qwen3.8-27B-abliteratedhuihui-ai27.8B≈18.1 GB at 4-bit
• Qwen3.5-27BAlibaba27.8B≈18.1 GB at 4-bit8 also selling it hosted
• Qwen3.5-27B-Claude-4.6-Opus-Reasoning-DistilledJackrong27.8B≈18.1 GB at 4-bit
• Qwen3.6 27BAlibaba27.8B≈18.1 GB at 4-bit12 also selling it hosted
• Qwen3.8 27BAlibaba27.8B≈18.1 GB at 4-bit12 also selling it hosted
• Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED-BF16AEON-727.8B≈18.1 GB at 4-bit
• Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAUDavidAU27.8B≈18.1 GB at 4-bit
• Qwen3.8-27B-UncensoredOrcaRouter27.8B≈18.1 GB at 4-bit
• Swift-Qwen3.8-27bukisai27.8B≈18.1 GB at 4-bit
• deepseek-vl2DeepSeek27.5B≈17.9 GB at 4-bit
• Gemma 3 27BGoogle27.4B≈17.8 GB at 4-bit11 also selling it hosted
• Gemma-3-R1984-27BVIDraft27.4B≈17.8 GB at 4-bit
• gemma-3-27b-it-abliteratedmlabonne27.4B≈17.8 GB at 4-bit
• Gemma 2 27BGoogle27.2B≈17.7 GB at 4-bit4 also selling it hosted
• datagemma-rag-27b-itGoogle27.2B≈17.7 GB at 4-bit
• medgemma-27b-text-itGoogle27.0B≈17.6 GB at 4-bit
• C2S-Scale-Gemma-2-27Bvandijklab27.0B≈17.6 GB at 4-bit
• Fara1.5-27BMicrosoft27.0B≈17.6 GB at 4-bit
• Gemma-SEA-LION-v4-27B-ITaisingapore27.0B≈17.6 GB at 4-bit2 also selling it hosted
• Huihui-Qwen3.5-27B-abliteratedhuihui-ai27.0B≈17.6 GB at 4-bit
• Qwen3.6-27B-AEON-Ultimate-Uncensored-BF16AEON-727.0B≈17.6 GB at 4-bit
• Qwen3.8-27B-DFlash2Z Lab27.0B≈17.6 GB at 4-bit
• Qwythos-27B-v1empero-ai27.0B≈17.6 GB at 4-bit
• harrier-oss-v1-27bMicrosoft27.0B≈17.6 GB at 4-bit
• thinkingcap-qwen3.6-27bsference27.0B≈17.6 GB at 4-bit1 also selling it hosted
• Gemma 4 26B A4B ITGoogle26.0B≈16.9 GB at 4-bit
• Gemma-4-26B-A4B-it-assistantGoogle26.0B≈16.9 GB at 4-bit
• diffusiongemma-26B-A4B-itGoogle25.8B≈16.8 GB at 4-bit
• Gemma 4 26B A4BGoogle25.8B≈16.8 GB at 4-bit7 also selling it hosted
• AriaRhymes.AI25.3B≈16.4 GB at 4-bit
• Voxtral Small 24B 2507Mistral AI24.3B≈15.8 GB at 4-bit4 also selling it hosted
• Dolphin Mistral 24B Venice EditionCognitive Computations24.0B≈15.6 GB at 4-bit1 also selling it hosted
• Mistral Small 3.1 24B Instruct 2503Mistral AI24.0B≈15.6 GB at 4-bit3 also selling it hosted
• Mistral-Small-3.2-24B-Instruct-2506Mistral AI24.0B≈15.6 GB at 4-bit1 also selling it hosted
• Mistral Small 3Mistral AI23.6B≈15.3 GB at 4-bit4 also selling it hosted
• sarvam-mSarvam AI23.6B≈15.3 GB at 4-bit
• solar-pro-preview-instructUpstage22.1B≈14.4 GB at 4-bit
• gpt-oss-safeguard-20bOpenAI21.5B≈14.0 GB at 4-bit6 also selling it hosted
• ERNIE-4.5-21B-A3B-PTBaidu21.0B≈13.7 GB at 4-bit1 also selling it hosted
• ERNIE-4.5-21B-A3B-ThinkingBaidu21.0B≈13.7 GB at 4-bit2 also selling it hosted
• GPT OSS 20BOpenAI20.9B≈13.6 GB at 4-bit20 also selling it hosted
• context-1chroma20.9B≈13.6 GB at 4-bit
• gpt-neox-20bEleutherAI20.7B≈13.5 GB at 4-bit
• FireRed-Image-Edit-1.0FireRed Team20.4B≈13.3 GB at 4-bit
• Qwen ImageAlibaba20.4B≈13.3 GB at 4-bit1 also selling it hosted
• Qwen-Image-2512Alibaba20.4B≈13.3 GB at 4-bit
• Qwen-Image-Edit-2509Alibaba20.4B≈13.3 GB at 4-bit
• Qwen-Image-Edit-2511Alibaba20.4B≈13.3 GB at 4-bit
• maple-previewdeepgrove20.2B≈13.1 GB at 4-bit
• Gemma-4-31B-JANG_4M-CRACKdealignai20.2B≈13.1 GB at 4-bit
• cogvlm2-llama3-chat-19Bzai-org19.5B≈12.7 GB at 4-bit
• LTX-2LTX.io18.9B≈12.3 GB at 4-bit
• cogvlm-chat-hfzai-org17.6B≈11.5 GB at 4-bit
• SenseNova-U1.5-8B-MoTsensenova17.5B≈11.4 GB at 4-bit
• Wan2.1-VACE-14BWan-AI17.3B≈11.3 GB at 4-bit
• HiDream-E1-1HiDream.ai17.1B≈11.1 GB at 4-bit
• HiDream-E1-FullHiDream.ai17.1B≈11.1 GB at 4-bit
• HiDream-I1-FullHiDream.ai17.1B≈11.1 GB at 4-bit
• Kimi-VL-A3B-ThinkingMoonshot AI16.4B≈10.7 GB at 4-bit
• Kimi-VL-A3B-Thinking-2506Moonshot AI16.4B≈10.7 GB at 4-bit
• Wan2.1-I2V-14B-480PWan-AI16.4B≈10.7 GB at 4-bit
• Wan2.1-I2V-14B-720PWan-AI16.4B≈10.7 GB at 4-bit
• LLaDA2.0-UniInclusionAI16.3B≈10.6 GB at 4-bit
• Wan2.2-S2V-14BWan-AI16.3B≈10.6 GB at 4-bit
• Ling-mini-2.0InclusionAI16.3B≈10.6 GB at 4-bit
• Ring-mini-2.0InclusionAI16.3B≈10.6 GB at 4-bit
• Instella-MoE-16B-A3B-Thinkamd16.0B≈10.4 GB at 4-bit
• deepseek-moe-16b-basedeepseek-ai16.0B≈10.4 GB at 4-bit
• deepseek-moe-16b-chatdeepseek-ai16.0B≈10.4 GB at 4-bit
• Moonlight-16B-A3B-InstructMoonshot AI16.0B≈10.4 GB at 4-bit
• starcoder2-15bbigcode16.0B≈10.4 GB at 4-bit1 also selling it hosted
• starcoderbigcode15.8B≈10.3 GB at 4-bit
• DeepSeek-Coder-V2-Lite-InstructDeepSeek15.7B≈10.2 GB at 4-bit1 also selling it hosted
• DeepSeek-V2-Litedeepseek-ai15.7B≈10.2 GB at 4-bit
• starchat-alphaHugging Face H415.5B≈10.1 GB at 4-bit
• Apriel-1.6-15b-ThinkerServiceNow-AI15.0B≈9.8 GB at 4-bit
• Phi-4-reasoning-vision-15BMicrosoft15.0B≈9.8 GB at 4-bit
• WizardCoder-15B-V1.0WizardLM Team15.0B≈9.8 GB at 4-bit
• starcoder2-15b-instruct-v0.1bigcode15.0B≈9.8 GB at 4-bit1 also selling it hosted
• Apriel-1.5-15b-ThinkerServiceNow-AI14.9B≈9.7 GB at 4-bit
• DeepCoder-14B-PreviewAgentica14.8B≈9.6 GB at 4-bit
• Qwen2.5-14B-Instruct-1MAlibaba14.8B≈9.6 GB at 4-bit
• Qwen2.5-Coder-14B-InstructAlibaba14.8B≈9.6 GB at 4-bit1 also selling it hosted
• Strand-Rust-Coder-14B-v1Fortytwo14.8B≈9.6 GB at 4-bit
• SuperNova-MediusArcee AI14.8B≈9.6 GB at 4-bit
• Qwen3 14BAlibaba14.8B≈9.6 GB at 4-bit8 also selling it hosted
• BAGEL-7B-MoTByteDance Seed14.7B≈9.5 GB at 4-bit