Updated 2026-10-21 · 159 open-weight models with adoption data
Trending open-weight models
Benchmarks tell you what a model can do; download counts tell you what the community actually runs. These are the open-weight models available via API, ranked by HuggingFace downloads — a proxy for real-world battle-testing, fine-tune ecosystem size, and how quickly bugs get found.
| # | |||||
|---|---|---|---|---|---|
| 1 | Google: Gemma 4 26B A4B google | 9.8M | 2K | $0.090 | 262K |
| 2 | Qwen: Qwen3 14Bqwen | 9.0M | 512 | $0.120 | 41K |
| 3 | Google: Gemma 4 31Bgoogle | 8.8M | 4K | $0.200 | 262K |
| 4 | Qwen: Qwen3.8 27Bqwen | 7.7M | 18K | $0.040 | 1M |
| 5 | Qwen: Qwen3.5-9Bqwen | 7.0M | 2K | $0.080 | 262K |
| 6 | Qwen: Qwen2.5 7B Instructqwen | 6.9M | 2K | $0.100 | 33K |
| 7 | Z.ai: GLM 5.3 Flash (batch)z-ai | 6.5M | 3K | $0.050 | 1.0M |
| 8 | Z.ai: GLM 5.3 Flashz-ai | 6.5M | 3K | $0.150 | 1.0M |
| 9 | OpenAI: gpt-oss-20b (batch)openai | 5.9M | 5K | $0.024 | 131K |
| 10 | OpenAI: gpt-oss-20bopenai | 5.9M | 5K | $0.018 | 131K |
| 11 | DeepSeek: DeepSeek V3.2deepseek | 5.5M | 2K | $0.280 | 164K |
| 12 | Meta: Llama 3.1 8B Instructmeta-llama | 5.3M | 9K | $0.050 | 131K |
| 13 | Qwen: Qwen3 VL 8B Instructqwen | 5.0M | 1K | $0.250 | 262K |
| 14 | Meta: Llama 3.2 1B Instructmeta-llama | 4.6M | 2K | $0.027 | 60K |
| 15 | DeepSeek: DeepSeek V4 Flash 0731deepseek | 4.2M | 4K | $0.440 | 1.0M |
| 16 | Qwen: Qwen3.6 35B A3Bqwen | 3.7M | 3K | $0.100 | 262K |
| 17 | OpenAI: gpt-oss-120b (batch)openai | 3.1M | 5K | $0.030 | 131K |
| 18 | OpenAI: gpt-oss-120bopenai | 3.1M | 5K | $0.150 | 131K |
| 19 | Qwen: Qwen3 32Bqwen | 2.8M | 769 | $0.080 | 131K |
| 20 | PrismML: Ternary Bonsai 2 27Bprism-ml | 2.7M | 3K | $0.075 | 262K |
| 21 | Qwen: Qwen3.8 Flashqwen | 2.6M | 6K | $0.150 | 1M |
| 22 | Qwen: Qwen3.6 27Bqwen | 2.4M | 2K | $0.250 | 262K |
| 23 | Z.ai: GLM 5.3 (batch)z-ai | 1.9M | 2K | $0.450 | 1.0M |
| 24 | Z.ai: GLM 5.3z-ai | 1.9M | 2K | $0.079 | 1.0M |
| 25 | Qwen: Qwen3.5-27Bqwen | 1.7M | 1K | $0.270 | 262K |
| 26 | DeepSeek: DeepSeek V4.1 Flash (batch)deepseek | 1.5M | 5K | $0.050 | 1.0M |
| 27 | DeepSeek: DeepSeek V4.1 Flashdeepseek | 1.5M | 5K | $0.049 | 1.0M |
| 28 | Qwen: Qwen3.5-35B-A3Bqwen | 1.4M | 2K | $0.140 | 262K |
| 29 | DeepSeek: DeepSeek V3deepseek | 1.4M | 4K | $0.257 | 164K |
| 30 | Thinking Machines: Inklingthinkingmachines | 1.4M | 2K | $1.00 | 524K |
| 31 | NVIDIA: Nemotron 3 Nano 30B A3Bnvidia | 1.2M | 831 | $0.060 | 262K |
| 32 | Meta: Llama 3.2 3B Instructmeta-llama | 1.1M | 3K | $0.051 | 131K |
| 33 | Qwen: Qwen3 30B A3Bqwen | 1.1M | 969 | $0.120 | 41K |
| 34 | Qwen: Qwen3 30B A3B Instruct 2507qwen | 1.1M | 849 | $0.090 | 262K |
| 35 | Google: Gemma 3 4Bgoogle | 1.1M | 2K | $0.040 | 131K |
| 36 | Z.ai: GLM 4.7 Flashz-ai | 1.1M | 2K | $0.060 | 200K |
| 37 | NVIDIA: Nemotron 3 Ultranvidia | 1.0M | 368 | $0.600 | 262K |
| 38 | MoonshotAI: Kimi K3 (batch)moonshotai | 989K | 12K | $2.28 | 1.0M |
| 39 | MoonshotAI: Kimi K3moonshotai | 989K | 12K | $0.239 | 1.0M |
| 40 | Z.ai: GLM 5z-ai | 946K | 2K | $1.00 | 205K |
| 41 | DeepSeek: DeepSeek V3 0324deepseek | 911K | 3K | $0.290 | 164K |
| 42 | MiniMax: MiniMax M2.5minimax | 827K | 2K | $0.270 | 205K |
| 43 | Thinking Machines: Inkling Smallthinkingmachines | 821K | 423 | $0.450 | 524K |
| 44 | DeepSeek: DeepSeek V4 Flash 0423deepseek | 808K | 2K | $0.091 | 1.0M |
| 45 | Z.ai: GLM 4.5 Airz-ai | 794K | 646 | $0.130 | 131K |
| 46 | Tencent: Hy3tencent | 769K | 984 | $0.132 | 262K |
| 47 | Qwen: Qwen3.5-122B-A10Bqwen | 722K | 640 | $0.260 | 262K |
| 48 | NVIDIA: Nemotron 3.5 Lightningnvidia | 719K | 243 | $0.070 | 262K |
| 49 | Qwen2.5 Coder 32B Instructqwen | 643K | 2K | $0.660 | 33K |
| 50 | Qwen: Qwen3.5 397B A17Bqwen | 629K | 2K | $0.390 | 262K |
| 51 | Z.ai: GLM 5.1z-ai | 578K | 2K | $1.40 | 205K |
| 52 | Qwen: Qwen3 Coder 30B A3B Instructqwen | 577K | 1K | $0.070 | 262K |
| 53 | Qwen: Qwen3 Coder Nextqwen | 521K | 2K | $0.120 | 262K |
| 54 | MiniMax: MiniMax M2.7minimax | 501K | 1K | $0.300 | 205K |
| 55 | Poolside: Laguna XS 2.1poolside | 461K | 257 | $0.060 | 262K |
| 56 | DeepSeek: DeepSeek V4 Flash Vision Expdeepseek | 415K | 956 | $0.216 | 1.0M |
| 57 | Poolside: Laguna S 2.1poolside | 386K | 1K | $0.090 | 1.0M |
| 58 | MoonshotAI: Kimi K2 0711moonshotai | 371K | 3K | $0.570 | 131K |
| 59 | Google: Gemma 3 27Bgoogle | 371K | 2K | $0.080 | 131K |
| 60 | Microsoft: Phi 4microsoft | 370K | 2K | $0.070 | 16K |
| 61 | Z.ai: GLM 5.2z-ai | 369K | 5K | $0.034 | 1.0M |
| 62 | MoonshotAI: Kimi K2.6moonshotai | 360K | 2K | $0.650 | 262K |
| 63 | DeepSeek: DeepSeek V4 Pro 0423deepseek | 358K | 6K | $0.955 | 1.0M |
| 64 | Google: Gemma 3 12Bgoogle | 352K | 880 | $0.050 | 131K |
| 65 | DeepSeek: DeepSeek V3.1deepseek | 332K | 861 | $0.250 | 164K |
| 66 | MoonshotAI: Kimi K2.5moonshotai | 313K | 3K | $0.500 | 262K |
| 67 | Qwen: Qwen3 VL 235B A22B Instructqwen | 294K | 426 | $0.210 | 262K |
| 68 | Mistral: Mistral Small 3.1 24Bmistralai | 274K | 1K | $0.351 | 128K |
| 69 | Qwen: Qwen3 VL 30B A3B Instructqwen | 273K | 619 | $0.150 | 262K |
| 70 | StepFun: Step 3.5 Flashstepfun | 269K | 840 | $0.100 | 262K |
| 71 | Qwen2.5 72B Instructqwen | 246K | 1K | $0.360 | 33K |
| 72 | Meta: Llama 3.3 70B Instructmeta-llama | 227K | 3K | $0.100 | 131K |
| 73 | Meta: Muse Glimmer 30Bmeta | 216K | 2K | $0.300 | 131K |
| 74 | MiniMax: MiniMax M3minimax | 209K | 2K | $0.300 | 1.0M |
| 75 | Mistral: Mistral Nemomistralai | 208K | 2K | $0.029 | 131K |
| 76 | Mistral: Mistral Small 3.2 24Bmistralai | 200K | 638 | $0.094 | 256K |
| 77 | ByteDance: UI-TARS 7B bytedance | 175K | 612 | $0.100 | 128K |
| 78 | DeepSeek: R1 0528deepseek | 172K | 2K | $0.500 | 164K |
| 79 | Qwen: Qwen3 Next 80B A3B Instructqwen | 171K | 1K | $0.090 | 262K |
| 80 | DeepSeek: DeepSeek V4 Pro 0813deepseek | 155K | 898 | $0.660 | 1.0M |
| 81 | Qwen: Qwen3 235B A22B Instruct 2507qwen | 152K | 811 | $0.090 | 262K |
| 82 | Meta: Llama 4 Scoutmeta-llama | 152K | 1K | $0.100 | 1.3M |
| 83 | Xiaomi: MiMo-V2.5xiaomi | 151K | 431 | $0.119 | 1.1M |
| 84 | MiniMax: MiniMax M2minimax | 133K | 2K | $0.255 | 205K |
| 85 | Mistral: Ministral 3 14B 2512mistralai | 125K | 345 | $0.200 | 262K |
| 86 | Mistral: Ministral 3 8B 2512 (batch)mistralai | 119K | 228 | $0.075 | 262K |
| 87 | Mistral: Ministral 3 8B 2512mistralai | 119K | 228 | $0.150 | 262K |
| 88 | Z.ai: GLM 4.5z-ai | 119K | 1K | $0.600 | 131K |
| 89 | MoonshotAI: Kimi K2.7 Codemoonshotai | 108K | 1K | $0.671 | 262K |
| 90 | Xiaomi: MiMo-V2.6-Flashxiaomi | 107K | 572 | $0.140 | 1.1M |
| 91 | Xiaomi: MiMo-V2.6-Proxiaomi | 104K | 758 | $0.435 | 1.1M |
| 92 | Mistral: Ministral 3 3B 2512mistralai | 98K | 308 | $0.100 | 131K |
| 93 | Tencent: Hy3 previewtencent | 90K | 313 | $0.180 | 262K |
| 94 | IBM: Granite 4.2 8Bibm-granite | 90K | 105 | $0.060 | 131K |
| 95 | Meta: Llama 3.1 70B Instructmeta-llama | 88K | 1K | $0.400 | 131K |
| 96 | Qwen: Qwen2.5 VL 72B Instructqwen | 86K | 661 | $0.800 | 128K |
| 97 | Z.ai: GLM 4.7z-ai | 84K | 2K | $0.540 | 205K |
| 98 | Moondream: Moondream 3 Previewmoondream | 82K | 692 | $0.300 | 32K |
| 99 | Mistral: Mistral Small 3mistralai | 73K | 985 | $0.050 | 33K |
| 100 | Tencent: Hunyuan A13B Instructtencent | 71K | 795 | $0.140 | 131K |
| 101 | Mistral: Mistral Small 4 (batch)mistralai | 71K | 454 | $0.075 | 262K |
| 102 | Mistral: Mistral Small 4mistralai | 71K | 454 | $0.150 | 262K |
| 103 | NVIDIA: Nemotron 3 Supernvidia | 60K | 291 | $0.300 | 262K |
| 104 | Qwen: Qwen3 Coder 480B A35Bqwen | 54K | 1K | $0.300 | 262K |
| 105 | MoonshotAI: Kimi K2 Thinkingmoonshotai | 49K | 2K | $0.600 | 262K |
| 106 | Mistral: Mixtral 8x22B Instructmistralai | 47K | 775 | $2.00 | 66K |
| 107 | Meta: Llama Guard 4 12Bmeta-llama | 47K | 133 | $0.180 | 164K |
| 108 | DeepSeek: DeepSeek V3.2 Expdeepseek | 45K | 1K | $0.270 | 164K |
| 109 | DeepSeek: DeepSeek V4.1 Flash 1017deepseek | 42K | 2K | $0.150 | 1.0M |
| 110 | Qwen: Qwen3 Next 80B A3B Thinkingqwen | 40K | 500 | $0.150 | 262K |
| 111 | MoonshotAI: Kimi K2 0905moonshotai | 39K | 790 | $0.600 | 262K |
| 112 | NVIDIA: Nemotron 3.5 Content Safetynvidia | 32K | 63 | $0.200 | 131K |
| 113 | Z.ai: GLM 4.5Vz-ai | 28K | 725 | $0.600 | 66K |
| 114 | Tencent: Hy-MT2-1.8Btencent | 28K | 1K | $0.044 | 8K |
| 115 | Qwen: Qwen3.8 2.4T A95Bqwen | 26K | 1K | $2.00 | 1.0M |
| 116 | Z.ai: GLM 4.6z-ai | 21K | 1K | $0.550 | 205K |
| 117 | OpenAI: gpt-oss-safeguard-20bopenai | 21K | 282 | $0.075 | 131K |
| 118 | Tencent: Hy4 previewtencent | 21K | 511 | $0.834 | 1.0M |
| 119 | Mistral: Voxtral Small 24B 2507mistralai | 20K | 545 | $0.100 | 33K |
| 120 | Google: Gemma 2 27Bgoogle | 20K | 577 | $0.650 | 8K |
| 121 | StepFun: Step 3.7 Flashstepfun | 18K | 462 | $0.200 | 262K |
| 122 | Qwen: Qwen3 235B A22B Thinking 2507qwen | 18K | 416 | $0.450 | 128K |
| 123 | Qwen: Qwen3 VL 30B A3B Thinkingqwen | 17K | 205 | $0.290 | 262K |
| 124 | LiquidAI: LFM2.5-24B-A2Bliquid | 17K | 533 | $0.050 | 128K |
| 125 | Xiaomi: MiMo-V2.5-Proxiaomi | 16K | 766 | $0.304 | 1.1M |
| 126 | DeepSeek: DeepSeek V3.1 Terminusdeepseek | 15K | 382 | $0.270 | 164K |
| 127 | Tencent: Hy-MT2-7Btencent | 14K | 213 | $0.074 | 8K |
| 128 | inclusionAI: Ling 3.0 Flash VLinclusionai | 14K | 124 | $0.060 | 131K |
| 129 | MiniMax: MiniMax M2.1minimax | 12K | 1K | $0.300 | 205K |
| 130 | Moondream: Moondream 3.1moondream | 11K | 89 | $0.300 | 32K |
| 131 | inclusionAI: Ling 3.0 Flashinclusionai | 10K | 434 | $0.021 | 262K |
| 132 | Meta: Llama 4 Maverickmeta-llama | 9K | 517 | $0.188 | 524K |
| 133 | IBM: Granite 4.0 Microibm-granite | 9K | 147 | $0.017 | 131K |
| 134 | Venice: Uncensoredcognitivecomputations | 8K | 757 | $0.200 | 128K |
| 135 | Mistral: Devstral 2 2512mistralai | 5K | 353 | $0.400 | 262K |
| 136 | TheDrummer: Cydonia 24B V4.1thedrummer | 4K | 83 | $0.300 | 131K |
| 137 | Z.ai: GLM 4.6Vz-ai | 4K | 398 | $0.300 | 131K |
| 138 | TheDrummer: Skyfall 36B V2thedrummer | 4K | 54 | $0.550 | 33K |
| 139 | Reflection: Beam 1.0 501B A23Breflection | 4K | 191 | $0.300 | 1.0M |
| 140 | Nex AGI: Nex-N2.5-Mininex-agi | 3K | 872 | $0.025 | 262K |
| 141 | Tencent: Hy-MT2-30B-A3Btencent | 3K | 509 | $0.074 | 8K |
| 142 | Meta: Muse Spark Openmeta | 3K | 617 | $1.25 | 1.0M |
| 143 | MiniMax: MiniMax-01minimax | 3K | 656 | $0.200 | 1.0M |
| 144 | Meituan: LongCat 2.0meituan | 3K | 282 | $0.300 | 1.0M |
| 145 | Arcee AI: Trinity Large Thinkingarcee-ai | 2K | 190 | $0.250 | 262K |
| 146 | Nous: Hermes 3 70B Instructnousresearch | 2K | 144 | $0.700 | 131K |
| 147 | StepFun: Step 5 Previewstepfun | 2K | 120 | $1.00 | 1.0M |
| 148 | Nex AGI: Nex-N2.5-Pronex-agi | 2K | 677 | $0.075 | 262K |
| 149 | TheDrummer: UnslopNemo 12Bthedrummer | 2K | 25 | $0.400 | 1.0M |
| 150 | Sao10K: Llama 3 8B Lunarissao10k | 2K | 147 | $0.040 | 8K |
| 151 | Cohere: Command Acohere | 2K | 398 | $2.50 | 256K |
| 152 | MythoMax 13Bgryphe | 1K | 409 | $0.080 | 8K |
| 153 | Nous: Hermes 3 405B Instructnousresearch | 1K | 276 | $1.00 | 131K |
| 154 | Nous: Hermes 4 405Bnousresearch | 1K | 97 | $1.00 | 131K |
| 155 | Reka Flash 3rekaai | 866 | 389 | $0.100 | 66K |
| 156 | Magnum v4 72Banthracite-org | 844 | 64 | $2.50 | 33K |
| 157 | Reka Edgerekaai | 806 | 134 | $0.100 | 16K |
| 158 | ReMM SLERP 13Bundi95 | 686 | 28 | $0.350 | 6K |
| 159 | Sao10K: Llama 3.1 Euryale 70B v2.2sao10k | 492 | 70 | $0.850 | 131K |
Download counts from the HuggingFace Hub API (2026-10-21); they include all quantized/fine-tuned derivatives of each repo family and are cumulative, not weekly. Prices verified 2026-10-21.
Why downloads matter as much as benchmarks
- Battle-testing: a model with 10M+ downloads has had its rough edges found and documented.
- Ecosystem: popular repos get more fine-tunes, quants, and tooling support.
- Longevity: high-adoption models rarely get abandoned by their maintainers.
Pair this with quality scores on the price-vs-performance chart, or check batch discounts to run them cheaper.