VRAM requirements for 285 open AI models
Memory needed at 8K context, weights + KV cache + overhead. Sorted by downloads. Type to filter.
| Model | Org | Params | Q4_K_M | Q8_0 | Attention | Max ctx | On HF |
|---|---|---|---|---|---|---|---|
| Qwen3 0.6B | Qwen | 752M | 1.9 GB | 2.2 GB | GQA | 40K | Apr 27, 2025 |
| Qwen3 14B | Qwen | 15B | 10.3 GB | 16.6 GB | GQA | 40K | Apr 27, 2025 |
| DeepSeek V3.2 | deepseek-ai | 685B MoE | 379 GB | 666 GB | MLA | 160K | Dec 1, 2025 |
| Qwen3 VL 8B Instruct | Qwen | 8.8B | 6.5 GB | 9.9 GB | GQA | 256K | Oct 11, 2025 |
| Qwen3.8 27B | Qwen | 28B | 16.8 GB | 28.4 GB | GQA | 256K | Aug 5, 2026 |
| Qwen2.5 7B Instruct | Qwen | 7.6B | 5.5 GB | 8.7 GB | GQA | 32K | Sep 16, 2024 |
| GLM 5.3 Flash | zai-org | 321B MoE | 184 GB | 318 GB | MLA | 1M | Aug 25, 2026 |
| gemma 4 31B it | 31B | 20.0 GB | 33.3 GB | GQA + sliding window | 256K | Mar 11, 2026 | |
| Qwen3.5 9B | Qwen | 9.7B | 6.4 GB | 10.0 GB | GQA | 256K | Feb 27, 2026 |
| Qwen3 8B | Qwen | 8.2B | 6.5 GB | 9.9 GB | GQA | 40K | Apr 27, 2025 |
| Unlimited OCR | baidu | 3.3B MoE | 2.9 GB | 4.0 GB | MHA | 32K | Jun 19, 2026 |
| gpt oss 20b | openai | 21B MoE | 11.8 GB | 12.3 GB | GQA + sliding window | 128K | Aug 4, 2025 |
| Qwen2.5 VL 7B Instruct | Qwen | 8.3B | 5.5 GB | 8.7 GB | GQA | 125K | Jan 26, 2025 |
| gemma 4 26B A4B it | 26B MoE | 17.2 GB | 26.4 GB | GQA + sliding window | 256K | Mar 11, 2026 | |
| Qwen2.5 1.5B Instruct | Qwen | 1.5B | 1.9 GB | 2.7 GB | GQA | 32K | Sep 17, 2024 |
| Qwen3.5 2B | Qwen | 2.3B | 2.2 GB | 2.8 GB | GQA | 256K | Feb 28, 2026 |
| Qwen3.5 4B | Qwen | 4.7B | 3.7 GB | 5.3 GB | GQA | 256K | Feb 27, 2026 |
| Llama 3.1 8B Instruct | meta-llama | 8.0B | 6.2 GB | 9.6 GB | GQA | 128K | Jul 18, 2024 |
| Qwen2.5 0.5B Instruct | Qwen | 494M | 1.2 GB | 1.4 GB | GQA | 32K | Sep 16, 2024 |
| Qwen3.6 35B A3B | Qwen | 36B MoE | 22.1 GB | 36.2 GB | GQA | 256K | Apr 15, 2026 |
| Qwen3 32B | Qwen | 33B | 21.1 GB | 35.1 GB | GQA | 40K | Apr 27, 2025 |
| Qwen3 VL 4B Instruct | Qwen | 4.4B | 4.1 GB | 5.8 GB | GQA | 256K | Oct 11, 2025 |
| Qwen3 4B | Qwen | 4.0B | 4.1 GB | 5.8 GB | GQA | 40K | Apr 27, 2025 |
| Qwen3 1.7B | Qwen | 2.0B | 2.6 GB | 3.3 GB | GQA | 40K | Apr 27, 2025 |
| Qwen3.8 Flash Next | Qwen | 180B MoE | 104 GB | 176 GB | GQA | 256K | Aug 24, 2026 |
| Qwen3.5 0.8B | Qwen | 873M | 1.5 GB | 1.7 GB | GQA | 256K | Feb 28, 2026 |
| Qwen3 4B Instruct 2507 | Qwen | 4.0B | 4.1 GB | 5.8 GB | GQA | 256K | Aug 5, 2025 |
| Qwen3.6 27B | Qwen | 28B | 17.0 GB | 28.0 GB | GQA | 256K | Apr 21, 2026 |
| gpt oss 120b | openai | 117B MoE | 59.5 GB | 60.1 GB | GQA + sliding window | 128K | Aug 4, 2025 |
| DeepSeek V4 Flash 0731 | deepseek-ai | 304B MoE | 175 GB | 302 GB | GQA | 1M | Jul 31, 2026 |
| Qwen2.5 3B Instruct | Qwen | 3.1B | 2.9 GB | 4.3 GB | GQA | 32K | Sep 17, 2024 |
| Qwen3 Coder 30B A3B Instruct | Qwen | 31B MoE | 18.7 GB | 31.7 GB | GQA | 256K | Jul 31, 2025 |
| Qwen2.5 VL 3B Instruct | Qwen | 3.8B | 2.8 GB | 4.0 GB | GQA | 125K | Jan 26, 2025 |
| DeepSeek OCR | deepseek-ai | 3.3B MoE | 2.9 GB | 4.0 GB | MHA | 8K | Oct 17, 2025 |
| GLM 5.3 | zai-org | 753B MoE | 430 GB | 748 GB | MLA | 1M | Aug 25, 2026 |
| Qwen2.5 Coder 14B Instruct | Qwen | 15B | 10.6 GB | 16.8 GB | GQA | 32K | Nov 6, 2024 |
| SmolVLM2 500M Video Instruct | HuggingFaceTB | 507M | 1.1 GB | 1.2 GB | GQA | 8K | Feb 11, 2025 |
| Qwen3.5 27B | Qwen | 28B | 17.0 GB | 28.0 GB | GQA | 256K | Feb 24, 2026 |
| DeepSeek V4.1 Flash | deepseek-ai | 763B MoE | 436 GB | 757 GB | GQA | 1M | Sep 10, 2026 |
| DeepSeek R1 | deepseek-ai | 684B MoE | 378 GB | 665 GB | MLA | 160K | Jan 20, 2025 |
| Qwen2.5 Coder 7B Instruct | Qwen | 7.6B | 5.5 GB | 8.7 GB | GQA | 32K | Sep 17, 2024 |
| Inkling | thinkingmachines | 952B MoE | 545 GB | 801 GB | GQA | — | Jul 14, 2026 |
| Qwen3.5 35B A3B | Qwen | 36B MoE | 21.5 GB | 35.4 GB | GQA | 256K | Feb 24, 2026 |
| DeepSeek V3 | deepseek-ai | 684B MoE | 378 GB | 665 GB | MLA | 160K | Dec 25, 2024 |
| Llama 3.2 1B Instruct | meta-llama | 1.2B | 1.6 GB | 2.1 GB | GQA | 128K | Sep 18, 2024 |
| DeepSeek V4 Flash | deepseek-ai | 291B MoE | 167 GB | 289 GB | GQA | 1M | Apr 22, 2026 |
| GLM 5 | zai-org | 754B MoE | 426 GB | 748 GB | MLA | 198K | Feb 11, 2026 |
| Mistral 7B Instruct v0.2 | mistralai | 7.2B | 47.7 GB | 8.6 GB | GQA | 32K | Dec 11, 2023 |
| GLM 4.7 Flash | zai-org | 31B MoE | 18.2 GB | 30.8 GB | MLA | 198K | Jan 19, 2026 |
| GLM 4.5 Air | zai-org | 110B MoE | 70.1 GB | 112 GB | GQA | 128K | Jul 20, 2025 |
| MiniMax M2.5 | MiniMaxAI | 229B MoE | 132 GB | 229 GB | GQA | 192K | Feb 12, 2026 |
| MiniCPM5 2B | openbmb | 2.5B | 2.5 GB | 3.5 GB | GQA | 128K | Sep 6, 2026 |
| gemma 3 4b it | 4.3B | 3.5 GB | 5.0 GB | GQA + sliding window | 128K | Feb 20, 2025 | |
| Hy3 | tencent | 299B MoE | 173 GB | 299 GB | GQA | 256K | Jul 2, 2026 |
| Llama 3.2 3B Instruct | meta-llama | 3.2B | 3.4 GB | 4.7 GB | GQA | 128K | Sep 18, 2024 |
| Kimi K3 | moonshotai | 2.78T MoE | 1584 GB | 2752 GB | MLA | 1M | Jun 13, 2026 |
| GLM 5.1 | zai-org | 754B MoE | 434 GB | 748 GB | MLA | 198K | Apr 3, 2026 |
| diffusiongemma 26B A4B it | 26B MoE | 17.1 GB | 26.4 GB | GQA + sliding window | 256K | Jun 9, 2026 | |
| HunyuanOCR | tencent | 1.1B | 2.0 GB | 2.5 GB | GQA | 128K | Nov 18, 2025 |
| Llama 3.2 1B | meta-llama | 1.2B | 1.8 GB | 2.4 GB | GQA | 128K | Sep 18, 2024 |
| Ornith 1.5 9B | ornith-ai | 9.7B | 6.6 GB | 10.0 GB | GQA | 256K | Aug 18, 2026 |
| Unlimited OCR 2 | baidu | 3.3B MoE | 3.0 GB | 4.0 GB | MHA | 128K | Oct 26, 2026 |
| DeepSeek R1 0528 Qwen3 8B | deepseek-ai | 8.2B | 6.5 GB | 9.9 GB | GQA | 128K | May 29, 2025 |
| DeepSeek V4.1 Flash 1017 | deepseek-ai | 763B MoE | 436 GB | 757 GB | GQA | 1M | Oct 17, 2026 |
| medgemma 4b it | 4.3B | 3.5 GB | 5.0 GB | GQA + sliding window | 128K | May 19, 2025 | |
| Mistral 7B v0.1 | mistralai | 7.2B | 5.5 GB | 8.6 GB | GQA | 32K | Sep 20, 2023 |
| Llama 3.1 8B | meta-llama | 8.0B | 6.2 GB | 9.6 GB | GQA | 128K | Jul 14, 2024 |
| DeepSeek R1 Distill Qwen 1.5B | deepseek-ai | 1.8B | 1.9 GB | 2.7 GB | GQA | 128K | Jan 20, 2025 |
| granite 4.1 3b | ibm-granite | 3.4B | 3.2 GB | 4.6 GB | GQA | 128K | Apr 6, 2026 |
| gemma 3 12b it | 12B | 8.5 GB | 13.4 GB | GQA + sliding window | 128K | Mar 1, 2025 | |
| Kimi K2.5 | moonshotai | 1.03T MoE | 580 GB | 1017 GB | MLA | 256K | Jan 1, 2026 |
| DeepSeek R1 Distill Qwen 32B | deepseek-ai | 33B | 21.2 GB | 35.1 GB | GQA | 128K | Jan 20, 2025 |
| Ornith 1.5 122B | ornith-ai | 125B MoE | 72.3 GB | 125 GB | GQA | 256K | Oct 16, 2026 |
| Llama 3.3 70B Instruct | meta-llama | 71B | 43.3 GB | 73.0 GB | GQA | 128K | Nov 26, 2024 |
| Hermes 3 Llama 3.1 8B | NousResearch | 8.0B | 6.2 GB | 9.6 GB | GQA | 128K | Jul 28, 2024 |
| Phi 3.5 mini instruct | microsoft | 3.8B | 5.7 GB | 7.2 GB | MHA | 128K | Aug 16, 2024 |
| Ornith 1.5 35B A3B | ornith-ai | 36B MoE | 21.4 GB | 36.2 GB | GQA | 256K | Aug 18, 2026 |
| Kimi K2.7 Code | moonshotai | 1.03T MoE | 579 GB | 1017 GB | MLA | 256K | Jun 11, 2026 |
| gemma 3 27b it | 27B | 17.3 GB | 28.7 GB | GQA + sliding window | 128K | Mar 1, 2025 | |
| Phi 3 mini 4k instruct | microsoft | 3.8B | 5.7 GB | 7.2 GB | MHA | 4K | Apr 22, 2024 |
| Llama 3.2 3B | meta-llama | 3.2B | 3.6 GB | 5.1 GB | GQA | 128K | Sep 18, 2024 |
| DeepSeek V4 Pro | deepseek-ai | 1.60T MoE | 891 GB | 1559 GB | GQA | 1M | Apr 22, 2026 |
| GLM 5.2 | zai-org | 753B MoE | 435 GB | 748 GB | MLA | 1M | Jun 16, 2026 |
| Phi 4 mini instruct | microsoft | 3.8B | 4.1 GB | 5.6 GB | GQA | 128K | Feb 19, 2025 |
| DeepSeek V3.1 | deepseek-ai | 685B MoE | 379 GB | 665 GB | MLA | 160K | Aug 21, 2025 |
| DeepSeek R1 Distill Qwen 7B | deepseek-ai | 7.6B | 5.5 GB | 8.7 GB | GQA | 128K | Jan 20, 2025 |
| gemma 3 4b pt | 4.3B | 3.6 GB | 5.4 GB | GQA + sliding window | 128K | Feb 20, 2025 | |
| Phi 3.5 vision instruct | microsoft | 4.1B | 5.8 GB | 7.6 GB | MHA | 128K | Aug 16, 2024 |
| Qwen3.5 9B Base | Qwen | 9.7B | 6.4 GB | 10.0 GB | GQA | 256K | Feb 26, 2026 |
| SmolVLM 500M Instruct | HuggingFaceTB | 507M | 1.1 GB | 1.2 GB | GQA | 8K | Jan 20, 2025 |
| DeepSeek OCR 2 | deepseek-ai | 3.4B MoE | 2.9 GB | 4.0 GB | MHA | 8K | Jan 27, 2026 |
| SmolLM2 1.7B Instruct | HuggingFaceTB | 1.7B | 3.0 GB | 3.7 GB | MHA | 8K | Oct 31, 2024 |
| Kimi K2 Instruct | moonshotai | 1.03T MoE | 579 GB | 1017 GB | MLA | 128K | Jul 11, 2025 |
| Phi 3 mini 128k instruct | microsoft | 3.8B | 5.7 GB | 7.2 GB | MHA | 128K | Apr 22, 2024 |
| medgemma 1.5 4b it | 4.3B | 3.5 GB | 5.0 GB | GQA + sliding window | 128K | Jan 7, 2026 | |
| UI TARS 1.5 7B | ByteDance-Seed | 8.3B | 5.5 GB | 8.7 GB | GQA | 125K | Apr 16, 2025 |
| SmolLM2 360M Instruct | HuggingFaceTB | 362M | 1.1 GB | 1.2 GB | GQA | 8K | Oct 31, 2024 |
| MiniMax M2.7 | MiniMaxAI | 229B MoE | 133 GB | 229 GB | GQA | 200K | Apr 9, 2026 |
| SmolLM3 3B | HuggingFaceTB | 3.1B | 3.0 GB | 4.3 GB | GQA | 64K | Jul 8, 2025 |
| phi 4 | microsoft | 15B | 10.4 GB | 16.7 GB | GQA | 16K | Dec 11, 2024 |
| Kimi K2.6 | moonshotai | 1.03T MoE | 586 GB | 1017 GB | MLA | 256K | Apr 14, 2026 |
| Apertus 8B Instruct 2509 | swiss-ai | 8.1B | 6.4 GB | 9.6 GB | GQA | 64K | Aug 13, 2025 |
| DeepSeek V4 Flash DSpark | deepseek-ai | 304B MoE | 175 GB | 302 GB | GQA | 1M | Jun 27, 2026 |
| Step 3.5 Flash | stepfun-ai | 199B | 116 GB | 199 GB | MHA + sliding window | 256K | Feb 1, 2026 |
| phi 2 | microsoft | 2.8B | 4.7 GB | 5.8 GB | MHA | 2K | Dec 13, 2023 |
| Llama Guard 4 12B | meta-llama | 12B | 9.1 GB | 14.2 GB | GQA | 10M | Apr 23, 2025 |
| LFM2.5 350M | LiquidAI | 354M | 0.8 GB | 1.0 GB | GQA | 125K | Mar 31, 2026 |
| Muse Glimmer 30B | meta-models | 30B | 17.9 GB | 28.5 GB | GQA + sliding window | 128K | Aug 9, 2026 |
| DeepSeek V3 0324 | deepseek-ai | 685B MoE | 378 GB | 665 GB | MLA | 160K | Mar 24, 2025 |
| DeepSeek R1 0528 | deepseek-ai | 685B MoE | 378 GB | 665 GB | MLA | 160K | May 28, 2025 |
| SmolLM2 360M | HuggingFaceTB | 362M | 1.1 GB | 1.2 GB | GQA | 8K | Oct 31, 2024 |
| Olmo 3 7B Instruct | allenai | 7.3B | 7.3 GB | 10.3 GB | MHA + sliding window | 64K | Nov 19, 2025 |
| Meta Llama 3.1 8B Instruct | NousResearch | 8.0B | 6.2 GB | 9.6 GB | GQA | 128K | Jul 24, 2024 |
| GOT OCR2 0 | stepfun-ai | 716M | 1.8 GB | 2.1 GB | MHA | 32K | Sep 12, 2024 |
| MiniMax M3 | MiniMaxAI | 427B MoE | 248 GB | 424 GB | GQA | 1M | Jun 2, 2026 |
| DeepSeek R1 Distill Qwen 14B | deepseek-ai | 15B | 10.6 GB | 16.8 GB | GQA | 128K | Jan 20, 2025 |
| falcon 7b | tiiuae | 7.2B | 9.6 GB | 12.1 GB | MHA | — | Apr 24, 2023 |
| GLM 4.5 | zai-org | 358B MoE | 205 GB | 358 GB | GQA | 128K | Jul 20, 2025 |
| OLMo 2 0425 1B | allenai | 1.5B | 2.5 GB | 3.1 GB | MHA | 4K | Apr 17, 2025 |
| DeepSeek V2 Lite | deepseek-ai | 16B MoE | 9.8 GB | 16.4 GB | MLA | 160K | May 15, 2024 |
| GOT OCR 2.0 hf | stepfun-ai | 561M | 1.8 GB | 2.0 GB | MHA | — | Nov 22, 2024 |
| granite 3.1 3b a800m instruct | ibm-granite | 3.3B MoE | 2.9 GB | 4.3 GB | GQA | 128K | Dec 6, 2024 |
| Llama 4 Scout 17B 16E Instruct | meta-llama | 109B MoE | 63.2 GB | 109 GB | GQA | 10M | Apr 2, 2025 |
| MiniMax M2 | MiniMaxAI | 229B MoE | 132 GB | 229 GB | GQA | 192K | Oct 22, 2025 |
| Mistral 7B Instruct v0.1 | mistralai | 7.2B | 5.5 GB | 8.6 GB | GQA | 32K | Sep 27, 2023 |
| LFM3 1.7B | LiquidAI | 1.7B | 1.8 GB | 2.5 GB | GQA + sliding window | 256K | Oct 23, 2026 |
| LFM2.5 1.2B Instruct | LiquidAI | 1.2B | 1.3 GB | 1.8 GB | GQA | 125K | Jan 6, 2026 |
| Olmo 3 1025 7B | allenai | 7.3B | 7.3 GB | 10.3 GB | MHA + sliding window | 64K | Sep 12, 2025 |
| granite 3.3 8b instruct | ibm-granite | 8.2B | 6.3 GB | 9.8 GB | GQA | 128K | Apr 9, 2025 |
| PaddleOCR VL 1.6 | PaddlePaddle | 959M | 1.3 GB | 1.7 GB | GQA | 128K | May 27, 2026 |
| Hy3 preview | tencent | 299B MoE | 173 GB | 299 GB | GQA | 256K | Apr 13, 2026 |
| deepseek coder 7b instruct v1.5 | deepseek-ai | 6.9B | 8.3 GB | 11.2 GB | MHA | 4K | Jan 25, 2024 |
| SmolLM3 3B Base | HuggingFaceTB | 3.1B | 3.0 GB | 4.3 GB | GQA | 64K | Jun 19, 2025 |
| Llama 2 7b hf | NousResearch | 6.7B | 8.3 GB | 11.1 GB | MHA | 4K | Jul 18, 2023 |
| DeepSeek Coder V2 Lite Instruct | deepseek-ai | 16B MoE | 10.5 GB | 16.4 GB | MLA | 160K | Jun 14, 2024 |
| ERNIE 4.5 21B A3B Thinking | baidu | 22B | 13.5 GB | 22.6 GB | GQA | 128K | Sep 8, 2025 |
| Meta Llama 3 8B Instruct | NousResearch | 8.0B | 6.2 GB | 9.6 GB | GQA | 8K | Apr 18, 2024 |
| granite vision 4.1 4b | ibm-granite | 4.0B | 3.2 GB | 4.6 GB | GQA | 128K | Apr 16, 2026 |
| granite 4.1 8b | ibm-granite | 8.8B | 6.8 GB | 10.5 GB | GQA | 128K | Apr 6, 2026 |
| olmOCR 2 7B 1025 | allenai | 8.3B | 5.5 GB | 8.7 GB | GQA | 125K | Oct 6, 2025 |
| MiMo V2.5 | XiaomiMiMo | 311B MoE | 180 GB | 308 GB | GQA | 1M | Apr 27, 2026 |
| DeepSeek R1 Distill Llama 8B | deepseek-ai | 8.0B | 6.2 GB | 9.6 GB | GQA | 128K | Jan 20, 2025 |
| Hunyuan A13B Instruct | tencent | 80B MoE | 47.1 GB | 81.2 GB | GQA | 32K | Jun 25, 2025 |
| deepseek vl2 tiny | deepseek-ai | 3.4B MoE | 3.0 GB | 4.5 GB | MHA | 4K | Dec 13, 2024 |
| MiniMax M3.1 Flash | MiniMaxAI | 220B MoE | 127 GB | 220 GB | GQA | 1M | Oct 17, 2026 |
| Olmo 3 7B Think | allenai | 7.3B | 7.3 GB | 10.3 GB | MHA + sliding window | 64K | Nov 18, 2025 |
| DeepSeek V2 Lite Chat | deepseek-ai | 16B MoE | 10.5 GB | 16.4 GB | MLA | 160K | May 15, 2024 |
| granite 4.2 8b | ibm-granite | 8.8B | 6.8 GB | 10.5 GB | GQA | 128K | Aug 7, 2026 |
| flint | RecursiveSI | 40B MoE | 23.6 GB | 40.3 GB | GQA + sliding window | 256K | Oct 21, 2026 |
| Kimi K2 Thinking | moonshotai | 1.03T MoE | 580 GB | 1017 GB | MLA | 256K | Nov 4, 2025 |
| DeepSeek V3.2 Exp | deepseek-ai | 685B MoE | 391 GB | 680 GB | MLA | 160K | Sep 29, 2025 |
| GLM 4.6V Flash | zai-org | 10B | 6.7 GB | 10.3 GB | GQA | 128K | Dec 7, 2025 |
| granite 3.2 8b instruct | ibm-granite | 8.2B | 6.3 GB | 9.8 GB | GQA | 128K | Feb 17, 2025 |
| Llama 3.2 11B Vision Instruct | meta-llama | 11B | 7.4 GB | 11.6 GB | GQA | 128K | Sep 18, 2024 |
| DeepSeek V4 Pro 0813 | deepseek-ai | 1.65T MoE | 941 GB | 1635 GB | GQA | 1M | Aug 13, 2026 |
| GLM 4.7 | zai-org | 358B MoE | 205 GB | 358 GB | GQA | 198K | Dec 22, 2025 |
| deepseek coder 6.7b base | deepseek-ai | 6.7B | 8.3 GB | 11.1 GB | MHA | 16K | Oct 23, 2023 |
| Kimi K3.1 Pro | moonshotai | 2.78T MoE | 1584 GB | 2753 GB | MLA | 1M | Nov 2, 2026 |
| internlm2 chat 7b | internlm | 7.7B | 6.0 GB | 9.2 GB | GQA | 32K | Jan 10, 2024 |
| Kimi Linear 48B A3B Instruct | moonshotai | 49B MoE | 28.9 GB | 49.6 GB | MLA | — | Oct 30, 2025 |
| internlm3 8b instruct | internlm | 8.8B | 6.0 GB | 9.7 GB | GQA | 32K | Jan 13, 2025 |
| Molmo2 8B | allenai | 8.7B | 6.5 GB | 9.9 GB | GQA | 36K | Dec 14, 2025 |
| OLMoE 1B 7B 0924 | allenai | 6.9B MoE | 5.4 GB | 8.4 GB | MHA | 4K | Jul 20, 2024 |
| step3 | stepfun-ai | 321B | 214 GB | 349 GB | MHA | 64K | Jul 28, 2025 |
| Seed OSS 36B Instruct | ByteDance-Seed | 36B | 23.0 GB | 38.5 GB | GQA | 512K | Aug 20, 2025 |
| SmolLM2 1.7B | HuggingFaceTB | 1.7B | 3.0 GB | 3.7 GB | MHA | 8K | Oct 30, 2024 |
| LFM2.5 VL 450M | LiquidAI | 449M | 0.8 GB | 1.0 GB | GQA | 125K | Apr 8, 2026 |
| MiMo V2.6 Flash RL | XiaomiMiMo | 311B MoE | 179 GB | 309 GB | GQA | 1M | Sep 21, 2026 |
| granite 4.2 3b | ibm-granite | 3.7B | 3.3 GB | 4.8 GB | GQA | 128K | Aug 7, 2026 |
| granite 3.3 2b instruct | ibm-granite | 2.5B | 2.6 GB | 3.6 GB | GQA | 128K | Apr 9, 2025 |
| Kimi VL A3B Instruct | moonshotai | 16B MoE | 10.8 GB | 16.8 GB | MLA | 128K | Apr 9, 2025 |
| GLM 4.1V 9B Thinking | zai-org | 10B | 6.7 GB | 10.3 GB | GQA | 64K | Jun 28, 2025 |
| olmOCR 7B 0225 preview | allenai | 8.3B | 5.5 GB | 8.7 GB | GQA | 32K | Jan 15, 2025 |
| granite 3.1 8b instruct | ibm-granite | 8.2B | 6.3 GB | 9.8 GB | GQA | 128K | Dec 6, 2024 |
| granite 4.2 30b | ibm-granite | 29B | 19.1 GB | 31.6 GB | GQA | 128K | Aug 7, 2026 |
| GLM Z1 32B 0414 | zai-org | 33B | 19.5 GB | 33.4 GB | GQA | 32K | Apr 8, 2025 |
| LFM2.5 2.6B | LiquidAI | 2.7B | 2.3 GB | 3.4 GB | GQA | 128K | Jul 28, 2026 |
| falcon 7b instruct | tiiuae | 7.2B | 9.4 GB | 11.8 GB | MHA | — | Apr 25, 2023 |
| DeepSeek V4 Flash Vision Exp | deepseek-ai | 305B MoE | 175 GB | 303 GB | GQA | 1M | Aug 31, 2026 |
| granite 4.0 1b | ibm-granite | 1.6B | 2.2 GB | 2.8 GB | GQA | 128K | Oct 7, 2025 |
| deepseek coder 1.3b instruct | deepseek-ai | 1.3B | 2.8 GB | 3.3 GB | MHA | 16K | Oct 29, 2023 |
| Moonlight 16B A3B Instruct | moonshotai | 16B MoE | 10.8 GB | 16.8 GB | MLA | 8K | Feb 22, 2025 |
| MiniMax M1 40k | MiniMaxAI | 456B MoE | 263 GB | 455 GB | GQA | 9.765625M | Jun 5, 2025 |
| LFM2.5 24B A2B | LiquidAI | 24B MoE | 14.3 GB | 24.5 GB | GQA | 128K | Oct 12, 2026 |
| MiMo 7B RL | XiaomiMiMo | 7.8B | 6.2 GB | 9.4 GB | GQA | 32K | Apr 29, 2025 |
| deepseek coder 6.7b instruct | deepseek-ai | 6.7B | 8.3 GB | 11.1 GB | MHA | 16K | Oct 29, 2023 |
| granite 4.0 h tiny | ibm-granite | 6.9B MoE | 4.6 GB | 7.5 GB | GQA | 128K | Sep 16, 2025 |
| Yi 34B Chat | 01-ai | 34B | 21.6 GB | 36.4 GB | GQA | 4K | Nov 22, 2023 |
| Qwen3.8 2.4T A95B | Qwen | 2.45T MoE | 1394 GB | 2423 GB | GQA | 256K | Aug 8, 2026 |
| gpt oss safeguard 20b | openai | 22B MoE | 11.8 GB | 12.3 GB | GQA + sliding window | 128K | Sep 18, 2025 |
| LLaDA2.1 mini | inclusionAI | 16B MoE | 10.3 GB | 17.1 GB | GQA | 32K | Feb 9, 2026 |
| Molmo2 4B | allenai | 4.9B | 4.3 GB | 6.2 GB | GQA | 36K | Dec 14, 2025 |
| GLM 4.6 | zai-org | 357B MoE | 204 GB | 357 GB | GQA | 198K | Sep 29, 2025 |
| granite 4.0 h small | ibm-granite | 32B MoE | 18.9 GB | 32.6 GB | GQA | 128K | Sep 16, 2025 |
| WeVisOne VL 4B | tencent | 4.8B | 4.3 GB | 6.2 GB | GQA | 256K | Oct 22, 2026 |
| DeepSeek V2 Chat | deepseek-ai | 236B MoE | 135 GB | 234 GB | MLA | 160K | Apr 28, 2024 |
| LFM3 VL 2B | LiquidAI | 1.9B | 1.8 GB | 2.5 GB | GQA + sliding window | 125K | Oct 19, 2026 |
| deepseek moe 16b base | deepseek-ai | 16B MoE | 12.5 GB | 18.6 GB | MHA | 4K | Jan 8, 2024 |
| Nemotron 3.5 Parse | nvidia | 631M | 1.1 GB | 1.5 GB | GQA | 16K | Sep 3, 2026 |
| MiMo V2.6 Distill Qwen 9B | XiaomiMiMo | 9.4B | 6.6 GB | 10.0 GB | GQA | 256K | Sep 21, 2026 |
| SmolLM 1.7B | HuggingFaceTB | 1.7B | 3.0 GB | 3.7 GB | MHA | 2K | Jul 14, 2024 |
| MiMo V2.6 Pro RL | XiaomiMiMo | 1.02T MoE | 587 GB | 1017 GB | GQA | 1M | Sep 21, 2026 |
| DeepSeek R1 Distill Llama 70B | deepseek-ai | 71B | 43.3 GB | 73.0 GB | GQA | 128K | Jan 20, 2025 |
| glm 4 9b | zai-org | 9.4B | 11.5 GB | 15.0 GB | MHA | 8K | Jun 4, 2024 |
| MiMo V2.6 Flash MOPD | XiaomiMiMo | 311B MoE | 179 GB | 309 GB | GQA | 1M | Sep 27, 2026 |
| Intern S1 | internlm | 241B MoE | 139 GB | 235 GB | GQA | 64K | Jul 24, 2025 |
| d1 3B | LiquidAI | 3.1B | 2.3 GB | 3.4 GB | GQA | 32K | Oct 5, 2026 |
| granite 4.0 micro | ibm-granite | 3.4B | 3.2 GB | 4.6 GB | GQA | 128K | Sep 16, 2025 |
| FastVLM 0.5B | apple | 759M | 1.2 GB | 1.5 GB | GQA | 32K | Aug 25, 2025 |
| Llama 3.1 70B | meta-llama | 71B | 42.7 GB | 73.0 GB | GQA | 128K | Jul 14, 2024 |
| gpt oss safeguard 120b | openai | 117B MoE | 59.5 GB | 60.1 GB | GQA + sliding window | 128K | Sep 18, 2025 |
| Qianfan OCR | baidu | 4.7B | 4.4 GB | 6.2 GB | GQA | 32K | Mar 18, 2026 |
| Yi 6B 200K | 01-ai | 6.1B | 4.4 GB | 7.0 GB | GQA | 195K | Nov 6, 2023 |
| LFM2.5 VL 3B | LiquidAI | 3.1B | 2.3 GB | 3.4 GB | GQA | 32K | Aug 11, 2026 |
| internlm2 5 7b chat | internlm | 7.7B | 6.0 GB | 9.2 GB | GQA | 32K | Jun 27, 2024 |
| SmolVLM Instruct | HuggingFaceTB | 2.2B | 3.0 GB | 3.8 GB | MHA | 16K | Nov 18, 2024 |
| Inkling 2 Small | thinkingmachines | 594B MoE | 346 GB | 590 GB | GQA | — | Nov 4, 2026 |
| ERNIE 4.5 21B A3B PT | baidu | 22B | 13.5 GB | 22.6 GB | GQA | 128K | Jun 28, 2025 |
| LFM2.5 8B A1B | LiquidAI | 8.5B MoE | 5.5 GB | 9.1 GB | GQA | 125K | May 28, 2026 |
| Yi 6B | 01-ai | 6.1B | 4.4 GB | 7.0 GB | GQA | 4K | Nov 1, 2023 |
| WeVisOne VL 2B | tencent | 2.4B | 2.8 GB | 3.6 GB | GQA | 256K | Oct 22, 2026 |
| MiMo V2 Flash | XiaomiMiMo | 310B MoE | 176 GB | 308 GB | GQA | 256K | Dec 16, 2025 |
| LFM2 350M | LiquidAI | 354M | 1.0 GB | 1.1 GB | GQA | 125K | Jul 10, 2025 |
| Yi 1.5 34B Chat | 01-ai | 34B | 21.6 GB | 36.4 GB | GQA | 4K | May 10, 2024 |
| GLM 4.5V | zai-org | 108B MoE | 61.4 GB | 108 GB | GQA | 64K | Aug 10, 2025 |
| LongCat Flash Chat | meituan-longcat | 562B MoE | 321 GB | 557 GB | MLA | 128K | Aug 29, 2025 |
| Muse Spark Open 816B A42B | meta-models | 822B MoE | 470 GB | 813 GB | GQA + sliding window | 1M | Sep 15, 2026 |
| LightOnOCR 3 4B | lightonai | 4.5B | 3.6 GB | 5.3 GB | GQA | 256K | Oct 6, 2026 |
| LightOnOCR 3 0.8B | lightonai | 853M | 1.5 GB | 1.7 GB | GQA | 256K | Oct 7, 2026 |
| d1 2B | LiquidAI | 1.9B | 1.8 GB | 2.5 GB | GQA + sliding window | 125K | Oct 24, 2026 |
| Kimi K2 Base | moonshotai | 1.03T MoE | 586 GB | 1017 GB | MLA | 128K | Jul 3, 2025 |
| Ling 3.0 tiny | inclusionAI | 7.9B MoE | 5.5 GB | 8.7 GB | MLA | 128K | Aug 10, 2026 |
| MiniMax M2.1 | MiniMaxAI | 229B MoE | 132 GB | 229 GB | GQA | 192K | Dec 20, 2025 |
| Llama 4 Maverick 17B 128E Instruct | meta-llama | 402B MoE | 228 GB | 399 GB | GQA | 1M | Apr 1, 2025 |
| Kimi VL A3B Thinking 2506 | moonshotai | 16B MoE | 10.8 GB | 16.8 GB | MLA | 128K | Jun 21, 2025 |
| LFM3 1.7B Base | LiquidAI | 1.7B | 1.7 GB | 2.5 GB | GQA + sliding window | 256K | Oct 26, 2026 |
| Yi 34B | 01-ai | 34B | 21.6 GB | 36.4 GB | GQA | 4K | Nov 1, 2023 |
| Trinity Mini | arcee-ai | 26B MoE | 15.8 GB | 26.9 GB | GQA + sliding window | 128K | Dec 1, 2025 |
| Hy4 preview | tencent | 780B MoE | 445 GB | 773 GB | MLA | 1M | Aug 27, 2026 |
| falcon 40b | tiiuae | 42B | 25.2 GB | 42.9 GB | GQA | — | May 24, 2023 |
| Nous Hermes 2 Mixtral 8x7B DPO | NousResearch | 47B MoE | 28.0 GB | 47.7 GB | GQA | 32K | Jan 11, 2024 |
| Yi 34B 200K | 01-ai | 34B | 21.6 GB | 36.4 GB | GQA | 195K | Nov 6, 2023 |
| Step 3.7 Flash | stepfun-ai | 201B | 119 GB | 203 GB | MHA + sliding window | 256K | May 23, 2026 |
| Nous Hermes 2 Yi 34B | NousResearch | 34B | 22.1 GB | 36.4 GB | GQA | 4K | Dec 23, 2023 |
| Kimi K2 Instruct 0905 | moonshotai | 1.03T MoE | 580 GB | 1017 GB | MLA | 256K | Sep 3, 2025 |
| Arete 27B | NousResearch | 27B | 17.4 GB | 28.1 GB | GQA | 256K | Nov 6, 2026 |
| Qwen AgentWorld 35B A3B | Qwen | 35B MoE | 21.6 GB | 35.4 GB | GQA | 256K | Jun 22, 2026 |
| MiMo V2.5 Pro | XiaomiMiMo | 1.02T MoE | 590 GB | 1017 GB | GQA | 1M | Apr 27, 2026 |
| Ling 3.0 flash | inclusionAI | 127B MoE | 73.5 GB | 127 GB | MLA | 256K | Aug 2, 2026 |
| Kimi VL A3B Thinking | moonshotai | 16B MoE | 10.3 GB | 17.2 GB | MLA | 128K | Apr 9, 2025 |
| North Mini Code 1.0 | CohereLabs | 30B MoE | 19.3 GB | 31.6 GB | GQA + sliding window | 488K | Jun 5, 2026 |
| GLM 4 32B 0414 | zai-org | 33B | 19.5 GB | 33.4 GB | GQA | 32K | Apr 7, 2025 |
| Hy Memory 30B A3B | tencent | 30B MoE | 18.5 GB | 31.1 GB | GQA | 256K | Oct 15, 2026 |
| Llama 3.2 11B Vision | meta-llama | 11B | 8.0 GB | 12.4 GB | GQA | 128K | Sep 18, 2024 |
| Ling 3.1 flash | inclusionAI | 572B MoE | 335 GB | 568 GB | MLA | 256K | Oct 13, 2026 |
| GLM 4.6V | zai-org | 108B MoE | 67.8 GB | 108 GB | GQA | 128K | Dec 7, 2025 |
| Qwen3.8 27B DSpine b16 | tencent | 2.0B | 2.2 GB | 3.1 GB | GQA + sliding window | 256K | Nov 7, 2026 |
| Intern Decision 4B | internlm | 4.5B | 3.6 GB | 5.3 GB | GQA | 256K | Sep 26, 2026 |
| Mage VL Cowork | microsoft | 4.7B | 4.3 GB | 6.2 GB | GQA | 256K | Oct 30, 2026 |
| Step3 VL 10B | stepfun-ai | 10B | 6.5 GB | 9.9 GB | GQA | 64K | Jan 13, 2026 |
| Ling 3.1 flash 1019 | inclusionAI | 572B MoE | 335 GB | 568 GB | MLA | 256K | Oct 27, 2026 |
| SenseNova Loop 9B Preview | sensenova | 9.7B | 6.4 GB | 10.0 GB | GQA | 256K | Oct 28, 2026 |
| Hermes 3 Llama 3.2 3B | NousResearch | 3.2B | 3.4 GB | 4.7 GB | GQA | 128K | Dec 3, 2024 |
| Ling 1T | inclusionAI | 1000B MoE | 566 GB | 993 GB | GQA | 32K | Oct 2, 2025 |
| ERNIE 5.5 Flash Base | baidu | 286B MoE | 164 GB | 284 GB | MLA | 1M | Oct 21, 2026 |
| Kimi Dev 72B | moonshotai | 73B | 44.6 GB | 75.1 GB | GQA | 128K | Jun 16, 2025 |
| MiniMax Text 01 | MiniMaxAI | 456B MoE | 263 GB | 455 GB | GQA | 9.765625M | Jan 12, 2025 |
| GLiDE 27B | fastino | 26B | 16.2 GB | 27.2 GB | GQA | 256K | Oct 19, 2026 |
| Falcon H1R 7B | tiiuae | 7.6B | 5.3 GB | 8.5 GB | GQA | 256K | Oct 29, 2025 |
| Hermes 4.3 36B | NousResearch | 36B | 23.0 GB | 38.5 GB | GQA | 512K | Nov 17, 2025 |
| LongCat 2.0 | meituan-longcat | 1.78T MoE | 1012 GB | 1758 GB | MLA | 256K | Jul 5, 2026 |
| Ling 3.0 flash Sante | inclusionAI | 127B MoE | 75.9 GB | 127 GB | MLA | 256K | Oct 19, 2026 |
| MiniMax H3 Image PE | MiniMaxAI | 27B | 16.8 GB | 28.0 GB | GQA | 256K | Nov 8, 2026 |
| UI TARS 7B DPO | ByteDance-Seed | 8.3B | 5.5 GB | 8.7 GB | GQA | 32K | Jan 22, 2025 |
| Groot 19B A3B 261023 | GrootTeam | 19B MoE | 12.7 GB | 19.9 GB | GQA | 256K | Oct 23, 2026 |
| granite 4.1 3b llmon preview | ibm-granite | 3.7B | 3.3 GB | 4.8 GB | GQA | 128K | Oct 1, 2026 |
| S5 nano | maritaca-ai | 4.7B | 3.7 GB | 5.4 GB | GQA | 256K | Nov 2, 2026 |
| Meta Llama Guard 2 8B | meta-llama | 8.0B | 6.2 GB | 9.6 GB | GQA | 8K | Apr 17, 2024 |
| Axion 1.0 35B | Axion1-0 | 35B MoE | 20.7 GB | 35.8 GB | GQA | 256K | Oct 16, 2026 |
| LongCat 2.5 | meituan-longcat | 1.78T MoE | 1014 GB | 1762 GB | MLA | 512K | Nov 4, 2026 |
| Llama 4 Scout 17B 16E | meta-llama | 109B MoE | 64.1 GB | 110 GB | GQA | 256K | Apr 2, 2025 |
| Raon 1 nano | KRAFTON | 3.5B | 3.0 GB | 4.3 GB | MLA | 256K | Oct 29, 2026 |
| SingVL S1 4B | inclusionAI | 4.5B | 3.6 GB | 5.3 GB | GQA | 256K | Nov 11, 2026 |
| Nex N2.6 | nex-agi | 321B MoE | 188 GB | 319 GB | MLA | 1M | Nov 10, 2026 |
| Maxwell V1 Flash | maxwell-model | 37B MoE | 22.1 GB | 37.6 GB | GQA | 512K | Oct 13, 2026 |