Home › Embed Embed a "Will it run on my GPU?" badge or widget For model cards, READMEs, blog posts and docs. Pick a model, then copy a snippet. Badges are static SVG files updated with the site. The widget lets readers choose their GPU or Mac and see the result. Both link back to RunsOnMy.
Model Qwen3 0.6B · runs on 8 GB+ VRAM Qwen3 14B · runs on 16 GB+ VRAM DeepSeek V3.2 · needs 379 GB VRAM Qwen3.8 27B · runs on 24 GB+ VRAM Qwen3 VL 8B Instruct · runs on 8 GB+ VRAM Qwen2.5 7B Instruct · runs on 8 GB+ VRAM GLM 5.3 Flash · needs 184 GB VRAM Gemma 4 31B IT · runs on 24 GB+ VRAM Qwen3.5 9B · runs on 8 GB+ VRAM Qwen3 8B · runs on 8 GB+ VRAM Unlimited OCR · runs on 8 GB+ VRAM gpt-oss 20B · runs on 16 GB+ VRAM Qwen3.5 2B · runs on 8 GB+ VRAM Qwen2.5 VL 7B Instruct · runs on 8 GB+ VRAM Gemma 4 26B A4B IT · runs on 24 GB+ VRAM Qwen2.5 1.5B Instruct · runs on 8 GB+ VRAM Qwen3.5 4B · runs on 8 GB+ VRAM Llama 3.1 8B Instruct · runs on 8 GB+ VRAM Qwen2.5 0.5B Instruct · runs on 8 GB+ VRAM Qwen3.6 35B A3B · runs on 32 GB+ VRAM Qwen3 32B · runs on 32 GB+ VRAM Qwen3 VL 4B Instruct · runs on 8 GB+ VRAM Qwen3 4B · runs on 8 GB+ VRAM Qwen3.8 Flash Next · needs 104 GB VRAM Qwen3.5 0.8B · runs on 8 GB+ VRAM Qwen3 1.7B · runs on 8 GB+ VRAM Qwen3 4B Instruct 2507 · runs on 8 GB+ VRAM Qwen3.6 27B · runs on 24 GB+ VRAM gpt-oss 120B · needs 60 GB VRAM DeepSeek V4 Flash 0731 · needs 175 GB VRAM Qwen3 Coder 30B A3B Instruct · runs on 24 GB+ VRAM Qwen2.5 3B Instruct · runs on 8 GB+ VRAM Qwen2.5 VL 3B Instruct · runs on 8 GB+ VRAM DeepSeek OCR · runs on 8 GB+ VRAM GLM 5.3 · needs 430 GB VRAM SmolVLM2 500M Video Instruct · runs on 8 GB+ VRAM Qwen3.5 27B · runs on 24 GB+ VRAM Qwen2.5 Coder 14B Instruct · runs on 16 GB+ VRAM DeepSeek R1 · needs 378 GB VRAM DeepSeek V4.1 Flash · needs 436 GB VRAM Qwen2.5 Coder 7B Instruct · runs on 8 GB+ VRAM Inkling · needs 545 GB VRAM Qwen3.5 35B A3B · runs on 32 GB+ VRAM DeepSeek V3 · needs 378 GB VRAM Llama 3.2 1B Instruct · runs on 8 GB+ VRAM DeepSeek V4 Flash · needs 167 GB VRAM GLM 5 · needs 426 GB VRAM GLM 4.7 Flash · runs on 24 GB+ VRAM Mistral 7B Instruct v0.2 · runs on 64 GB+ VRAM GLM 4.5 Air · needs 70 GB VRAM MiniMax M2.5 · needs 132 GB VRAM Gemma 3 4B IT · runs on 8 GB+ VRAM Hy3 · needs 173 GB VRAM MiniCPM5 2B · runs on 8 GB+ VRAM Llama 3.2 3B Instruct · runs on 8 GB+ VRAM Kimi K3 · needs 1584 GB VRAM GLM 5.1 · needs 434 GB VRAM Llama 3.2 1B · runs on 8 GB+ VRAM DiffusionGemma 26B A4B IT · runs on 24 GB+ VRAM HunyuanOCR · runs on 8 GB+ VRAM Unlimited OCR 2 · runs on 8 GB+ VRAM Ornith 1.5 9B · runs on 8 GB+ VRAM Cosmos Reason2 2B · runs on 8 GB+ VRAM DeepSeek R1 0528 Qwen3 8B · runs on 8 GB+ VRAM DeepSeek V4.1 Flash 1017 · needs 436 GB VRAM MedGemma 4B IT · runs on 8 GB+ VRAM DeepSeek R1 Distill Qwen 1.5B · runs on 8 GB+ VRAM Mistral 7B v0.1 · runs on 8 GB+ VRAM Llama 3.1 8B · runs on 8 GB+ VRAM Granite 4.1 3B · runs on 8 GB+ VRAM Gemma 3 12B IT · runs on 12 GB+ VRAM Kimi K2.5 · needs 580 GB VRAM DeepSeek R1 Distill Qwen 32B · runs on 32 GB+ VRAM Llama 3.3 70B Instruct · runs on 64 GB+ VRAM Hermes 3 Llama 3.1 8B · runs on 8 GB+ VRAM Ornith 1.5 35B A3B · runs on 32 GB+ VRAM Phi 3.5 Mini Instruct · runs on 8 GB+ VRAM Kimi K2.7 Code · needs 579 GB VRAM Llama 3.2 3B · runs on 8 GB+ VRAM Phi 3 Mini 4K Instruct · runs on 8 GB+ VRAM DeepSeek V4 Pro · needs 891 GB VRAM Gemma 3 27B IT · runs on 24 GB+ VRAM DeepSeek V3.1 · needs 379 GB VRAM GLM 5.2 · needs 435 GB VRAM Phi 4 Mini Instruct · runs on 8 GB+ VRAM DeepSeek R1 Distill Qwen 7B · runs on 8 GB+ VRAM Gemma 3 4B PT · runs on 8 GB+ VRAM SmolVLM 500M Instruct · runs on 8 GB+ VRAM Phi 3.5 Vision Instruct · runs on 8 GB+ VRAM SmolLM2 1.7B Instruct · runs on 8 GB+ VRAM Kimi K2 Instruct · needs 579 GB VRAM DeepSeek OCR 2 · runs on 8 GB+ VRAM Phi 3 Mini 128K Instruct · runs on 8 GB+ VRAM MedGemma 1.5 4B IT · runs on 8 GB+ VRAM UI TARS 1.5 7B · runs on 8 GB+ VRAM SmolLM2 360M Instruct · runs on 8 GB+ VRAM MiniMax M2.7 · needs 133 GB VRAM Phi 4 · runs on 16 GB+ VRAM Kimi K2.6 · needs 586 GB VRAM SmolLM3 3B · runs on 8 GB+ VRAM DeepSeek V4 Flash DSpark · needs 175 GB VRAM Step 3.5 Flash · needs 116 GB VRAM Llama Guard 4 12B · runs on 12 GB+ VRAM Apertus 8B Instruct 2509 · runs on 8 GB+ VRAM Phi 2 · runs on 8 GB+ VRAM LFM2.5 350M · runs on 8 GB+ VRAM Muse Glimmer 30B · runs on 24 GB+ VRAM DeepSeek V3 0324 · needs 378 GB VRAM DeepSeek R1 0528 · needs 378 GB VRAM SmolLM2 360M · runs on 8 GB+ VRAM Olmo 3 7B Instruct · runs on 12 GB+ VRAM Meta Llama 3.1 8B Instruct · runs on 8 GB+ VRAM GOT OCR2.0 · runs on 8 GB+ VRAM MiniMax M3 · needs 248 GB VRAM GLM 4.5 · needs 205 GB VRAM Falcon 7B · runs on 12 GB+ VRAM GOT OCR 2.0 HF · runs on 8 GB+ VRAM OLMo 2 0425 1B · runs on 8 GB+ VRAM DeepSeek V2 Lite · runs on 12 GB+ VRAM Granite 3.1 3B A800M Instruct · runs on 8 GB+ VRAM DeepSeek R1 Distill Qwen 14B · runs on 16 GB+ VRAM MiniMax M2 · needs 132 GB VRAM Mistral 7B Instruct v0.1 · runs on 8 GB+ VRAM Llama 4 Scout 17B 16E Instruct · needs 63 GB VRAM LFM3 1.7B · runs on 8 GB+ VRAM LFM2.5 1.2B Instruct · runs on 8 GB+ VRAM Olmo 3 1025 7B · runs on 12 GB+ VRAM Granite 3.3 8B Instruct · runs on 8 GB+ VRAM Hy3 Preview · needs 173 GB VRAM PaddleOCR VL 1.6 · runs on 8 GB+ VRAM SmolLM3 3B Base · runs on 8 GB+ VRAM ERNIE 4.5 21B A3B Thinking · runs on 16 GB+ VRAM Llama 2 7B HF · runs on 12 GB+ VRAM Llama 3.1 70B Instruct · runs on 64 GB+ VRAM DeepSeek Coder V2 Lite Instruct · runs on 16 GB+ VRAM Meta Llama 3 8B Instruct · runs on 8 GB+ VRAM DeepSeek Coder 7B Instruct v1.5 · runs on 12 GB+ VRAM olmOCR 2 7B 1025 · runs on 8 GB+ VRAM Granite 4.1 8B · runs on 8 GB+ VRAM Granite Vision 4.1 4B · runs on 8 GB+ VRAM MiMo V2.5 · needs 180 GB VRAM Hunyuan A13B Instruct · runs on 64 GB+ VRAM DeepSeek R1 Distill Llama 8B · runs on 8 GB+ VRAM MiniMax M3.1 Flash · needs 127 GB VRAM DeepSeek V2 Lite Chat · runs on 16 GB+ VRAM Olmo 3 7B Think · runs on 12 GB+ VRAM Granite 4.2 8B · runs on 8 GB+ VRAM Flint · runs on 32 GB+ VRAM Kimi K2 Thinking · needs 580 GB VRAM DeepSeek V3.2 Exp · needs 391 GB VRAM GLM 4.6V Flash · runs on 8 GB+ VRAM Kimi K3.1 Pro · needs 1584 GB VRAM GLM 4.7 · needs 205 GB VRAM Llama 3.2 11B Vision Instruct · runs on 12 GB+ VRAM InternLM2 Chat 7B · runs on 8 GB+ VRAM Granite 3.2 8B Instruct · runs on 8 GB+ VRAM Llama Guard 3 8B · runs on 8 GB+ VRAM Meta Llama 3 70B Instruct · runs on 64 GB+ VRAM InternLM3 8B Instruct · runs on 8 GB+ VRAM DeepSeek Coder 6.7B Base · runs on 12 GB+ VRAM Kimi Linear 48B A3B Instruct · runs on 48 GB+ VRAM Molmo2 8B · runs on 8 GB+ VRAM DeepSeek V4 Pro 0813 · needs 941 GB VRAM step3 · needs 214 GB VRAM OLMoE 1B 7B 0924 · runs on 8 GB+ VRAM Seed OSS 36B Instruct · runs on 32 GB+ VRAM Gemma 4 31B · runs on 32 GB+ VRAM LFM2.5 VL 450M · runs on 8 GB+ VRAM SmolLM2 1.7B · runs on 8 GB+ VRAM MiMo V2.6 Flash RL · needs 179 GB VRAM Granite 4.2 3B · runs on 8 GB+ VRAM Granite 3.3 2B Instruct · runs on 8 GB+ VRAM GLM 4.1V 9B Thinking · runs on 8 GB+ VRAM Granite 3.1 8B Instruct · runs on 8 GB+ VRAM olmOCR 7B 0225 Preview · runs on 8 GB+ VRAM Falcon 7B Instruct · runs on 12 GB+ VRAM DeepSeek VL2 Tiny · runs on 8 GB+ VRAM Kimi VL A3B Instruct · runs on 16 GB+ VRAM Granite 4.2 30B · runs on 24 GB+ VRAM GLM Z1 32B 0414 · runs on 24 GB+ VRAM LFM2.5 2.6B · runs on 8 GB+ VRAM Granite 4.0 1B · runs on 8 GB+ VRAM DeepSeek Coder 1.3B Instruct · runs on 8 GB+ VRAM Moonlight 16B A3B Instruct · runs on 16 GB+ VRAM DeepSeek V4 Flash Vision Exp · needs 175 GB VRAM MiniMax M1 40K · needs 263 GB VRAM DeepSeek Coder 6.7B Instruct · runs on 12 GB+ VRAM WeVisOne VL 4B · runs on 8 GB+ VRAM Granite 4.0 H Tiny · runs on 8 GB+ VRAM Tiny Aya Global · runs on 8 GB+ VRAM MiMo 7B RL · runs on 8 GB+ VRAM LFM2.5 24B A2B · runs on 24 GB+ VRAM Yi 34B Chat · runs on 32 GB+ VRAM gpt-oss Safeguard 20B · runs on 16 GB+ VRAM Qwen3.8 2.4T A95B · needs 1394 GB VRAM Molmo2 4B · runs on 8 GB+ VRAM LLaDA2.1 Mini · runs on 16 GB+ VRAM GLM 4.6 · needs 204 GB VRAM MiMo V2.6 Distill Qwen 9B · runs on 8 GB+ VRAM DeepSeek V2 Chat · needs 135 GB VRAM LFM3 VL 2B · runs on 8 GB+ VRAM Granite 4.0 H Small · runs on 24 GB+ VRAM DeepSeek MoE 16B Base · runs on 16 GB+ VRAM Aya Expanse 8B · runs on 8 GB+ VRAM Nemotron 3.5 Parse · runs on 8 GB+ VRAM Intern S1 · needs 139 GB VRAM Sage A1.5 · runs on 32 GB+ VRAM MiMo V2.6 Pro RL · needs 587 GB VRAM SmolLM 1.7B · runs on 8 GB+ VRAM MiMo V2.6 Flash MOPD · needs 179 GB VRAM DeepSeek R1 Distill Llama 70B · runs on 64 GB+ VRAM GLM 4 9B · runs on 16 GB+ VRAM d1 3B · runs on 8 GB+ VRAM Granite 4.0 Micro · runs on 8 GB+ VRAM FastVLM 0.5B · runs on 8 GB+ VRAM LFM2.5 VL 3B · runs on 8 GB+ VRAM Llama 3.1 70B · runs on 64 GB+ VRAM InternLM2.5 7B Chat · runs on 8 GB+ VRAM gpt-oss Safeguard 120B · needs 60 GB VRAM Yi 6B 200K · runs on 8 GB+ VRAM Inkling 2 Small · needs 346 GB VRAM ERNIE 4.5 21B A3B PT · runs on 16 GB+ VRAM WeVisOne VL 2B · runs on 8 GB+ VRAM SmolVLM Instruct · runs on 8 GB+ VRAM LFM2.5 8B A1B · runs on 8 GB+ VRAM Yi 6B · runs on 8 GB+ VRAM LFM2 1.2B · runs on 8 GB+ VRAM Qianfan OCR · runs on 8 GB+ VRAM MiMo V2 Flash · needs 176 GB VRAM Yi 1.5 34B Chat · runs on 32 GB+ VRAM GLM 4.5V · needs 61 GB VRAM LongCat Flash Chat · needs 321 GB VRAM d1 2B · runs on 8 GB+ VRAM LightOnOCR 3 0.8B · runs on 8 GB+ VRAM Muse Spark Open 816B A42B · needs 470 GB VRAM LightOnOCR 3 4B · runs on 8 GB+ VRAM Nous Hermes 2 Mixtral 8x7B DPO · runs on 48 GB+ VRAM Kimi K2 Base · needs 586 GB VRAM MiniMax M2.1 · needs 132 GB VRAM Ling 3.0 Tiny · runs on 8 GB+ VRAM Llama 4 Maverick 17B 128E Instruct · needs 228 GB VRAM Kimi VL A3B Thinking 2506 · runs on 16 GB+ VRAM LFM3 1.7B Base · runs on 8 GB+ VRAM Yi 34B · runs on 32 GB+ VRAM Olmo 3 32B Think · runs on 24 GB+ VRAM Trinity Mini · runs on 24 GB+ VRAM Falcon 40B · runs on 32 GB+ VRAM Hy4 Preview · needs 445 GB VRAM Arete 27B · runs on 24 GB+ VRAM Intern S1 Pro · needs 524 GB VRAM TranslateGemma 4B IT · runs on 8 GB+ VRAM Yi 34B 200K · runs on 32 GB+ VRAM Kimi K2 Instruct 0905 · needs 580 GB VRAM Step 3.7 Flash · needs 119 GB VRAM Nous Hermes 2 Yi 34B · runs on 32 GB+ VRAM Apertus v1.5 8B · runs on 8 GB+ VRAM Qwen AgentWorld 35B A3B · runs on 32 GB+ VRAM MiMo V2.5 Pro · needs 590 GB VRAM Kimi VL A3B Thinking · runs on 16 GB+ VRAM Ling 3.0 Flash · needs 74 GB VRAM GLM 4 32B 0414 · runs on 24 GB+ VRAM North Mini Code 1.0 · runs on 24 GB+ VRAM Apertus v1.5 70B · runs on 64 GB+ VRAM Llama 3.1 405B Instruct · needs 236 GB VRAM Ling 3.0 Flash VL · needs 74 GB VRAM Hy Memory 30B A3B · runs on 24 GB+ VRAM Llama 3.2 11B Vision · runs on 12 GB+ VRAM Ling 3.1 Flash · needs 335 GB VRAM GLM 4.6V · needs 68 GB VRAM Intern Decision 4B · runs on 8 GB+ VRAM Mage VL Cowork · runs on 8 GB+ VRAM Step3 VL 10B · runs on 8 GB+ VRAM Raon 1 Mini Preview · runs on 24 GB+ VRAM Ling 3.1 Flash 1019 · needs 335 GB VRAM SenseNova Loop 9B Preview · runs on 8 GB+ VRAM Ling 1T · needs 566 GB VRAM ERNIE 5.5 Flash Base · needs 164 GB VRAM Hermes 3 Llama 3.2 3B · runs on 8 GB+ VRAM Kimi Dev 72B · runs on 64 GB+ VRAM Ling 2.6 Flash · needs 62 GB VRAM GLiDE 27B · runs on 24 GB+ VRAM MiniMax Text 01 · needs 263 GB VRAM Falcon H1R 7B · runs on 8 GB+ VRAM Hermes 4.3 36B · runs on 32 GB+ VRAM MiniMax H3 Image PE · runs on 24 GB+ VRAM Ling 3.0 Flash Sante · needs 76 GB VRAM UI TARS 7B DPO · runs on 8 GB+ VRAM Granite 4.1 3B Llmon Preview · runs on 8 GB+ VRAM LongCat 2.5 · needs 1014 GB VRAM MiniMax M1 80K · needs 263 GB VRAM Meta Llama Guard 2 8B · runs on 8 GB+ VRAM Raon 1 Nano · runs on 8 GB+ VRAM SingVL S1 4B · runs on 8 GB+ VRAM Nex N2.6 · needs 188 GB VRAM Granite 4.1 8B Llmon Preview · runs on 8 GB+ VRAM Maxwell V1 Flash · runs on 32 GB+ VRAM KORMo Web 11B · runs on 12 GB+ VRAM SenseNova Loop 9B · runs on 8 GB+ VRAM
Badge
The smallest discrete-GPU VRAM size where the model fits at 4-bit (Q4_K_M) with an 8K context, using the published formula . Models too big for 64 GB show the memory they need instead.
Markdown (GitHub, Hugging Face model cards) [](https://runsonmy.com/model/qwen3-8b/) HTML <a href="https://runsonmy.com/model/qwen3-8b/"><img src="https://runsonmy.com/badge/qwen3-8b.svg" alt="Qwen3 8B: runs on 8 GB+ VRAM (RunsOnMy)"></a> Widget
Script (recommended; the link inside still works without JavaScript) <div class="runsonmy-widget" data-model="qwen3-8b"><a href="https://runsonmy.com/model/qwen3-8b/">Will Qwen3 8B run on my GPU? Check on RunsOnMy</a></div>
<script async src="https://runsonmy.com/embed.js"></script> Plain iframe <iframe src="https://runsonmy.com/embed/m/qwen3-8b/" width="100%" height="220" style="max-width:440px;border:0" loading="lazy" title="Will Qwen3 8B run on my GPU? (RunsOnMy)"></iframe> Options: data-gpu="rtx-4090" preselects a card (slugs are in /api/gpus.json ), and data-theme="dark" switches to a dark style. For the iframe, use ?gpu=…&theme=dark.
Privacy and fair use No cookies, no fingerprinting and no third-party requests. The widget page loads nothing except itself, and every verdict is precomputed. Embeds don't load our analytics, so your readers aren't counted or tracked. We only see a visit if a reader clicks through to runsonmy.com. Free to use under CC BY 4.0 . Please keep the RunsOnMy link, which is the attribution. Numbers are estimates from published formulas, not benchmarks. If one looks wrong, write to corrections . Examples