Popular Open-Weight Models for Local Deployment
| Version | Vendor | License | |||||||
|---|---|---|---|---|---|---|---|---|---|
 Hy4 preview | Tencent | Aug 28, 2026 | 770 | 49 | 798.2est. | 467.3 | 286.1est. | 235.4 | Apache-2.0 |
 GLM 5.3 | Z.ai / Zhipu AI | Aug 25, 2026 | 744 | 40 | 801.4 | 467.3 | 253.9 | 228.5 | GLM-5.3 License |
| GLM 5.3 Flash | Z.ai / Zhipu AI | Aug 25, 2026 | 320 | 18 | 341 | 199.7 | 108.7 | 97.6 | MIT |
 Qwen3.8 Flash Next | Alibaba | Aug 24, 2026 | 176 | 6 | 188.2 | 111.3 | 78.9 | 74.5 | Qwen Community License 1.0 |
 DeepSeek V4 Pro | DeepSeek | Aug 13, 2026 | 1,600 | 49 | 873.4 | 849.7 | 594.6est. | 375.5est. | MIT |
| Qwen3.8 2.4T A95B | Alibaba | Aug 8, 2026 | 2,446 | 95 | 2,600.2 | 1,310.9 | 730.7 | 564 | Qwen3.8-Max License |
| Qwen3.8 27B | Alibaba | Aug 5, 2026 | 27.8 | 27.8 | 29 | 16.5 | 9.8 | 6.7 | Apache-2.0 |
| DeepSeek V4 Flash | DeepSeek | Jul 31, 2026 | 284 | 13 | 161.9 | 155.1 | 96.8 | 86.9 | MIT |
| Hy3 | Tencent | Jul 6, 2026 | 295 | 21 | 317.7 | 182.2 | 107.2 | 71.1 | Apache-2.0 |
| GLM 5.2 | Z.ai / Zhipu AI | Jun 16, 2026 | 744 | 40 | 801.4 | 465.8 | 253.9 | 228.5 | MIT |
 Kimi K3 | Moonshot AI | Jun 13, 2026 | 2,800 | 104 | 1,561.2 | 1,508.7 | 861.3 | 648.9 | Kimi K3 License |
| Kimi K2.7 Code | Moonshot AI | Jun 12, 2026 | 1,000 | 32 | 594.5 | 583.7 | 339.5 | 303.9 | Modified MIT |
 Gemma 4 12B | Google DeepMind | May 23, 2026 | 12 | 12 | 12.7 | 7.1 | 4.7 | 2.8est. | Apache-2.0 |
| Qwen3.6 27B | Alibaba | Apr 21, 2026 | 27.8 | 27.8 | 28.6 | 16.8 | 11.8 | 6.5est. | Apache-2.0 |
| Qwen3.6 35B A3B | Alibaba | Apr 15, 2026 | 36 | 3 | 36.9 | 22.1 | 12.3 | 10 | Apache-2.0 |
| Kimi K2.6 | Moonshot AI | Apr 14, 2026 | 1,000 | 32 | 594.5 | 583.7 | 340.5 | 234.7est. | Modified MIT |
| Gemma 4 31B | Google DeepMind | Mar 11, 2026 | 31.3 | 31.3 | 32.6 | 18.3 | 11.8 | 7.3est. | Apache-2.0 |
| Gemma 4 26B A4B | Google DeepMind | Mar 11, 2026 | 25.8 | 4 | 26.9 | 16.9 | 10.5 | 6.1est. | Apache-2.0 |
| Gemma 4 E4B | Google DeepMind | Mar 2, 2026 | 8 | 8 | 8.2 | 5 | 3.8 | 1.9est. | Apache-2.0 |
| Qwen3.5 9B | Alibaba | Feb 27, 2026 | 9.7 | 9.7 | 9.5 | 5.7 | 4.1 | 2.3est. | Apache-2.0 |
| GLM 4.7 Flash | Z.ai / Zhipu AI | Jan 19, 2026 | 30 | 3 | 31.8 | 18.3 | 11.9 | 9.8 | MIT |
| Kimi K2.5 | Moonshot AI | Jan 1, 2026 | 1,000 | 32 | 1,036.6est. | 557.4est. | 371.6est. | 234.7est. | Modified MIT |
| DeepSeek V3.2 | DeepSeek | Dec 1, 2025 | 671 | 37 | 695.5est. | 374est. | 249.3est. | 157.5est. | MIT |
 Ministral 3 14B | Mistral AI | Oct 31, 2025 | 14 | 14 | 14.4 | 8.2 | 5.5 | 3.7 | Apache-2.0 |
| Ministral 3 8B | Mistral AI | Oct 31, 2025 | 8.9 | 8.9 | 9 | 5.2 | 3.6 | 2.4 | Apache-2.0 |
| Kimi Linear 48B A3B | Moonshot AI | Oct 30, 2025 | 49.1 | 3 | 52.2 | 30.1 | 17.9 | 10.9 | MIT |
| gpt-oss Safeguard 20B | OpenAI | Sep 18, 2025 | 20.9 | 3.6 | 12.1 | 11.6 | 7.8est. | 4.9est. | Apache-2.0 |
| Kimi K2 | Moonshot AI | Sep 3, 2025 | 1,000 | 32 | 1,036.6est. | 557.4est. | 371.6est. | 234.7est. | Modified MIT |
| gpt-oss 120B | OpenAI | Aug 4, 2025 | 116.8 | 5.1 | 63.4 | 62.8 | 43.4est. | 27.4est. | Apache-2.0 |
| gpt-oss 20B | OpenAI | Aug 4, 2025 | 20.9 | 3.6 | 12.1 | 11.6 | 7.8est. | 4.9est. | Apache-2.0 |
| GLM 4.5 Air | Z.ai / Zhipu AI | Jul 20, 2025 | 106 | 12 | 117.5 | 73 | 47.4 | 40.1 | MIT |
| Mistral Small 3.2 24B | Mistral AI | Jun 19, 2025 | 24 | 24 | 25.1 | 14.3 | 9.5 | 5.6est. | Apache-2.0 |
| DeepSeek R1 Qwen3 8B | DeepSeek | May 29, 2025 | 8.2 | 8.2 | 8.7 | 5 | 3.5 | 2.4 | MIT |
| Qwen3 32B | Alibaba | Apr 27, 2025 | 32.8 | 32.8 | 34.8 | 19.8 | 12.8 | 8.3 | Apache-2.0 |
| Qwen3 8B | Alibaba | Apr 27, 2025 | 8.2 | 8.2 | 8.7 | 5 | 3.5 | 2.4 | Apache-2.0 |
 Llama 4 Scout 17B 16E | Meta AI | Apr 2, 2025 | 108.6 | 17 | 114.5 | 65.4 | 42.4 | 35 | Llama 4 Community License |
| DeepSeek R1 Distill Qwen 32B | DeepSeek | Jan 20, 2025 | 32.8 | 32.8 | 34.8 | 19.9 | 13.1 | 7.7est. | MIT |
| DeepSeek R1 Distill Qwen 14B | DeepSeek | Jan 20, 2025 | 14.8 | 14.8 | 15.7 | 9 | 6.5 | 3.5est. | MIT |
| Llama 3.3 70B | Meta AI | Nov 26, 2024 | 70.6 | 70.6 | 75 | 42.5 | 27.4 | 16.8 | Llama 3.3 Community License |
| Llama 3.2 3B | Meta AI | Sep 18, 2024 | 3.2 | 3.2 | 3.4 | 2 | 1.2est. | 0.8est. | Llama 3.2 Community License |
| Llama 3.1 8B | Meta AI | Jul 18, 2024 | 8 | 8 | 8.5 | 4.9 | 3.7 | 1.9est. | Llama 3.1 Community License |
| Mistral Nemo | Mistral AI | Jul 17, 2024 | 12.3 | 12.3 | 13 | 7.5 | 5.4 | 2.9est. | Apache-2.0 |
| Mistral 7B v0.3 | Mistral AI | May 22, 2024 | 7.3 | 7.3 | 7.7 | 4.4 | 2.5 | 1.8 | Apache-2.0 |
Each size is the GGUF file of that bit width in the repository (Q8_0 for 8-bit, unsloth dynamic quants preferred for the rest); missing tiers are estimated from bytes per parameter and tagged · Prices are median asking prices, not sold prices · Methodology