Can I run it? Pick a model
Current models (newest first)
| Model | On HF since | Params | Q4_K_M needs ~ | Q8_0 needs ~ |
|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017 (MoE) | 2026-10-17 | 763.2B | 486.8 GB | 852.4 GB |
| LFM2.5 24B A2B (MoE) | 2026-10-12 | 24.0B | 15.9 GB | 27.4 GB |
| MiMo V2.6 Flash RL (MoE) | 2026-09-21 | 310.8B | 199.0 GB | 347.9 GB |
| Muse Spark Open 816B (MoE) | 2026-09-15 | 822.0B | 524.5 GB | 918.3 GB |
| DeepSeek V4.1 Flash (MoE) | 2026-09-10 | 763.2B | 486.8 GB | 852.4 GB |
| GLM 5.3 Flash (MoE) | 2026-08-25 | 321.3B | 205.2 GB | 359.1 GB |
| GLM 5.3 (MoE) | 2026-08-25 | 753.3B | 480.5 GB | 841.4 GB |
| Qwen3.8 Flash Next (MoE) | 2026-08-24 | 180.0B | 115.3 GB | 201.5 GB |
| granite 4.2 8b | 2026-08-07 | 8.8B | 6.9 GB | 11.1 GB |
| granite 4.2 30b | 2026-08-07 | 29.3B | 20.3 GB | 34.3 GB |
| Qwen3.8 27B | 2026-08-05 | 27.8B | 18.6 GB | 31.9 GB |
| NVIDIA Nemotron 3.5 Lightning 30B A3B (MoE) | 2026-08-01 | 31.6B | 20.7 GB | 35.9 GB |
| Hy3 (MoE) | 2026-07-02 | 298.8B | 192.1 GB | 335.3 GB |
| Kimi K3 (MoE) | 2026-06-13 | 2779.9B | 1770.6 GB | 3102.4 GB |
| MiniMax M3 (MoE) | 2026-06-02 | 427.0B | 272.9 GB | 477.5 GB |
| LFM2.5 8B A1B (MoE) | 2026-05-28 | 8.5B | 6.0 GB | 10.1 GB |
| Gemma 4 12B | 2026-05-23 | 12.0B | 8.8 GB | 14.5 GB |
| DeepSeek V4 Flash (MoE) | 2026-04-22 | 290.9B | 186.2 GB | 325.5 GB |
| DeepSeek V4 Pro (MoE) | 2026-04-22 | 1598.8B | 1018.9 GB | 1784.8 GB |
| Qwen3.6 27B | 2026-04-21 | 27.8B | 18.6 GB | 31.9 GB |
| Qwen3.6 35B A3B (MoE) | 2026-04-15 | 36.0B | 23.6 GB | 40.8 GB |
| Kimi K2.6 (MoE) | 2026-04-14 | 1026.9B | 654.6 GB | 1146.5 GB |
| MiniMax M2.7 (MoE) | 2026-04-09 | 228.7B | 147.2 GB | 256.8 GB |
| Mistral Medium 3.5 128B | 2026-03-31 | 127.7B | 83.4 GB | 144.5 GB |
| Gemma 4 31B | 2026-03-11 | 31.3B | 22.0 GB | 37.0 GB |
| Gemma 4 26B A4B (MoE) | 2026-03-11 | 25.8B | 17.4 GB | 29.8 GB |
| NVIDIA Nemotron 3 Super 120B A12B (MoE) | 2026-03-10 | 123.6B | 79.3 GB | 138.5 GB |
| NVIDIA Nemotron 3 Nano 4B | 2026-03-07 | 4.0B | 3.2 GB | 5.1 GB |
| Gemma 4 E4B | 2026-03-02 | 8.0B | 5.8 GB | 9.6 GB |
| Gemma 4 E2B | 2026-03-02 | 5.1B | 3.9 GB | 6.4 GB |
| Qwen3.5 2B | 2026-02-28 | 2.3B | 2.1 GB | 3.2 GB |
| Qwen3.5 9B | 2026-02-27 | 9.7B | 6.9 GB | 11.5 GB |
| Qwen3.5 4B | 2026-02-27 | 4.7B | 3.7 GB | 5.9 GB |
| Qwen3.5 27B | 2026-02-24 | 27.8B | 18.6 GB | 31.9 GB |
| Qwen3.5 35B A3B (MoE) | 2026-02-24 | 36.0B | 23.6 GB | 40.8 GB |
| Qwen3 Coder Next (MoE) | 2026-01-30 | 79.7B | 51.4 GB | 89.6 GB |
| Mistral Small 4 119B (2603) (MoE) | 2026-01-23 | 119.4B | 76.7 GB | 133.9 GB |
| Phi 4 reasoning vision 15B | 2026-01-23 | 15.1B | 11.1 GB | 18.3 GB |
| GLM 4.7 Flash | 2026-01-19 | 31.2B | 20.7 GB | 35.7 GB |
| Olmo 3.1 32B | 2025-12-10 | 32.2B | 22.2 GB | 37.6 GB |
| GLM 4.6V Flash | 2025-12-07 | 10.3B | 7.3 GB | 12.3 GB |
| NVIDIA Nemotron 3 Nano 30B A3B (MoE) | 2025-12-04 | 31.6B | 20.7 GB | 35.9 GB |
| DeepSeek V3.2 (MoE) | 2025-12-01 | 685.4B | 437.2 GB | 765.5 GB |
| Devstral Small 2 24B (2512) | 2025-11-28 | 24.0B | 16.6 GB | 28.1 GB |
| Ministral 3 14B (2512) | 2025-10-31 | 13.9B | 10.1 GB | 16.8 GB |
| Ministral 3 8B (2512) | 2025-10-31 | 8.9B | 6.8 GB | 11.1 GB |
| Ministral 3 3B (2512) | 2025-10-31 | 3.8B | 3.5 GB | 5.3 GB |
| Kimi Linear 48B A3B (MoE) | 2025-10-30 | 49.1B | 32.0 GB | 55.5 GB |
| Qwen3 VL 8B | 2025-10-11 | 8.8B | 6.8 GB | 11.0 GB |
| Qwen3 4B (2507) | 2025-08-05 | 4.0B | 3.8 GB | 5.7 GB |
| gpt-oss 20B (MoE) | 2025-08-04 | 20.9B | 14.0 GB | 24.0 GB |
| gpt oss 120b (MoE) | 2025-08-04 | 116.8B | 75.1 GB | 131.1 GB |
| Qwen3 Coder 30B A3B (MoE) | 2025-07-31 | 30.5B | 20.4 GB | 35.1 GB |
| Qwen3 30B A3B (2507) (MoE) | 2025-07-28 | 30.5B | 20.4 GB | 35.1 GB |
| GLM 4.5 Air (MoE) | 2025-07-20 | 110.5B | 71.7 GB | 124.6 GB |
| Llama 4 Scout 17B 16E (MoE) | 2025-04-02 | 108.6B | 70.6 GB | 122.6 GB |
Older models
Kept for reference; hidden from the checker unless you tick "Show older models".
| Model | On HF since | Params | Q4_K_M needs ~ | Q8_0 needs ~ |
|---|---|---|---|---|
| Qwen3 8B | 2025-04-27 | 8.2B | 6.4 GB | 10.3 GB |
| Qwen3 14B | 2025-04-27 | 14.8B | 10.7 GB | 17.7 GB |
| Qwen3 32B | 2025-04-27 | 32.8B | 22.5 GB | 38.2 GB |
| Qwen3 30B-A3B (MoE) | 2025-04-27 | 30.5B | 20.4 GB | 35.1 GB |
| Mistral Small 3.1 24B | 2025-03-11 | 24.0B | 16.6 GB | 28.1 GB |
| Gemma 3 12B | 2025-03-01 | 12.2B | 9.0 GB | 14.8 GB |
| Gemma 3 27B | 2025-03-01 | 27.4B | 18.8 GB | 32.0 GB |
| Gemma 3 4B | 2025-02-20 | 4.3B | 3.5 GB | 5.6 GB |
| DeepSeek R1 Distill Qwen 7B | 2025-01-20 | 7.6B | 5.7 GB | 9.3 GB |
| DeepSeek R1 Distill Qwen 14B | 2025-01-20 | 14.8B | 10.8 GB | 17.9 GB |
| DeepSeek R1 Distill Qwen 32B | 2025-01-20 | 32.8B | 22.5 GB | 38.2 GB |
| DeepSeek R1 Distill Llama 70B | 2025-01-20 | 70.6B | 46.9 GB | 80.7 GB |
| Phi-4 14B | 2024-12-11 | 14.7B | 10.8 GB | 17.8 GB |
| Llama 3.3 70B Instruct | 2024-11-26 | 70.6B | 46.9 GB | 80.7 GB |
| Qwen2.5 Coder 32B Instruct | 2024-11-06 | 32.8B | 22.5 GB | 38.2 GB |
| Qwen2.5 14B Instruct | 2024-09-16 | 14.8B | 10.8 GB | 17.9 GB |
| Llama 3.1 8B Instruct | 2024-07-18 | 8.0B | 6.2 GB | 10.1 GB |
| Mistral 7B Instruct v0.3 | 2024-05-22 | 7.2B | 5.8 GB | 9.2 GB |
| Llama 2 7B | 2023-07-13 | 6.7B | 7.0 GB | 10.3 GB |
Memory values are estimates (weights + 10% + 1 GB). Dates are when the model repo was created on Hugging Face.
Get the weekly "new models & what it takes to run them" email
Free. One email a week. Unsubscribe anytime.