{"name":"Twoody local LLM calculator data","version":"2026-10-16.1","reviewed_on":"2026-10-16","catalogue_verified_on":"2026-10-15","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","url":"https://www.twoody.com/local-llm-calculator","method":{"memory":"As Twoody's app computes it: the model's file + its vision file and 0.3 GiB + the KV cache at the window Twoody opens the model with (q8_0 for a hybrid model, f16 otherwise) + sliding windows + recurrent state + 0.3 GiB of work buffers.","work_buffers_bytes":322122547.2,"vision_buffers_bytes":322122547.2,"q8_0_factor":0.53125,"reference_model":"LLaMA 2 7B Q4_0 (3.8226 GB)","gpu_efficiency":0.85,"cpu_efficiency":0.55,"mac_gpu_share":{"up to 32 GB":0.667,"36 to 128 GB":0.75,"above 128 GB":0.8},"document_tokens":14000},"sources":{"apple":"https://www.apple.com/newsroom/","apple_specs":"https://support.apple.com/specs","llama_cpp_apple":"https://github.com/ggml-org/llama.cpp/discussions/4167","llama_cpp_cuda":"https://github.com/ggml-org/llama.cpp/discussions/15013","llama_cpp_vulkan":"https://github.com/ggml-org/llama.cpp/discussions/10879","llama_cpp_gpt_oss":"https://github.com/ggml-org/llama.cpp/discussions/15396","nvidia":"https://www.nvidia.com/en-us/geforce/graphics-cards/compare/","hugging_face":"https://huggingface.co/models","reading_speed":"https://www.sciencedirect.com/science/article/abs/pii/S0749596X19300786"},"devices":[{"name":"M5 Max","kind":"apple","memory":[36,48,64,128],"vram":null,"bandwidth":614,"tg":119.92,"pp":3219.99,"measured":true,"old_build":false,"id":"m5-max","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[{"bandwidth":460,"tg":93,"pp":2580,"measured":false,"memory":[36]}]},{"name":"M5 Ultra","kind":"apple","memory":[96,256,512],"vram":null,"bandwidth":1200,"tg":150,"pp":4900,"measured":false,"old_build":false,"id":"m5-ultra","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M5 Pro","kind":"apple","memory":[24,48,64],"vram":null,"bandwidth":307,"tg":66.33,"pp":1340.44,"measured":true,"old_build":false,"id":"m5-pro","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M6","kind":"apple","memory":[16,24,32],"vram":null,"bandwidth":170,"tg":34,"pp":940,"measured":false,"old_build":false,"id":"m6","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[{"bandwidth":153,"tg":32,"memory":[16]}]},{"name":"M5","kind":"apple","memory":[16,24,32],"vram":null,"bandwidth":153,"tg":31.88,"pp":722.79,"measured":true,"old_build":false,"id":"m5","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"A18 Pro","kind":"apple","memory":[8],"vram":null,"bandwidth":60,"tg":12.5,"pp":85,"measured":false,"old_build":false,"id":"a18-pro","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M3 Ultra","kind":"apple","memory":[96,256,512],"vram":null,"bandwidth":819,"tg":88.4,"pp":1073.09,"measured":true,"old_build":true,"id":"m3-ultra","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M4 Max","kind":"apple","memory":[36,48,64,128],"vram":null,"bandwidth":546,"tg":83.06,"pp":885.68,"measured":true,"old_build":true,"id":"m4-max","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[{"bandwidth":410,"tg":69.95,"pp":713.93,"memory":[36]}]},{"name":"M4 Pro","kind":"apple","memory":[24,48,64],"vram":null,"bandwidth":273,"tg":49.64,"pp":364.06,"measured":true,"old_build":true,"id":"m4-pro","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M4","kind":"apple","memory":[16,24,32],"vram":null,"bandwidth":120,"tg":24.11,"pp":221.29,"measured":true,"old_build":true,"id":"m4","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M3 Max","kind":"apple","memory":[36,48,64,96,128],"vram":null,"bandwidth":400,"tg":66.31,"pp":759.7,"measured":true,"old_build":true,"id":"m3-max","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[{"bandwidth":300,"tg":56.58,"pp":567.59,"memory":[36,96]}]},{"name":"M3 Pro","kind":"apple","memory":[18,36],"vram":null,"bandwidth":150,"tg":30.65,"pp":269.49,"measured":true,"old_build":true,"id":"m3-pro","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M3","kind":"apple","memory":[8,16,24],"vram":null,"bandwidth":100,"tg":21.34,"pp":186.75,"measured":true,"old_build":true,"id":"m3","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M2 Ultra","kind":"apple","memory":[64,128,192],"vram":null,"bandwidth":800,"tg":88.64,"pp":1013.81,"measured":true,"old_build":true,"id":"m2-ultra","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M2 Max","kind":"apple","memory":[32,64,96],"vram":null,"bandwidth":400,"tg":60.99,"pp":537.6,"measured":true,"old_build":true,"id":"m2-max","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M2 Pro","kind":"apple","memory":[16,32],"vram":null,"bandwidth":200,"tg":37.87,"pp":294.24,"measured":true,"old_build":true,"id":"m2-pro","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M2","kind":"apple","memory":[8,16,24],"vram":null,"bandwidth":100,"tg":21.7,"pp":145.91,"measured":true,"old_build":true,"id":"m2","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M1 Ultra","kind":"apple","memory":[64,128],"vram":null,"bandwidth":800,"tg":74.93,"pp":772.24,"measured":true,"old_build":true,"id":"m1-ultra","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M1 Max","kind":"apple","memory":[32,64],"vram":null,"bandwidth":400,"tg":54.61,"pp":400.26,"measured":true,"old_build":true,"id":"m1-max","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M1 Pro","kind":"apple","memory":[16,32],"vram":null,"bandwidth":200,"tg":35.52,"pp":232.55,"measured":true,"old_build":true,"id":"m1-pro","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"M1","kind":"apple","memory":[8,16],"vram":null,"bandwidth":68.25,"tg":14.15,"pp":107.81,"measured":true,"old_build":true,"id":"m1","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"Intel MacBook Air, MacBook Pro 13″ (2018–2020)","kind":"intel_mac","memory":[8,16,32],"vram":null,"bandwidth":45,"tg":6.371517386519415,"pp":40.5,"measured":false,"old_build":false,"id":"intel-macbook","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"Intel MacBook Pro 15″, 16″ (2018–2019)","kind":"intel_mac","memory":[16,32,64],"vram":null,"bandwidth":40,"tg":5.673612543841551,"pp":36.0,"measured":false,"old_build":false,"id":"intel-macbook-pro-16","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"Intel iMac, Mac mini (2017–2020)","kind":"intel_mac","memory":[8,16,32,64],"vram":null,"bandwidth":42.7,"tg":6.050788232073558,"pp":38.43000000000001,"measured":false,"old_build":false,"id":"intel-imac","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"iMac Pro, Mac Pro (2017–2019)","kind":"intel_mac","memory":[32,64,128],"vram":null,"bandwidth":100,"tg":13.888538168228076,"pp":90.0,"measured":false,"old_build":false,"id":"intel-imac-pro","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"GeForce RTX 5090 (32 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":32,"bandwidth":1792,"tg":300.4,"pp":14970.15,"measured":true,"old_build":false,"id":"rtx-5090","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 5080 (16 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":16,"bandwidth":960,"tg":184.68,"pp":9487.7,"measured":true,"old_build":false,"id":"rtx-5080","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 5070 Ti (16 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":16,"bandwidth":896,"tg":182.43,"pp":8419.56,"measured":true,"old_build":false,"id":"rtx-5070-ti","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 5070 (12 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":12,"bandwidth":672,"tg":128.21,"pp":5783.44,"measured":true,"old_build":false,"id":"rtx-5070","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 5060 Ti (16 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":16,"bandwidth":448,"tg":93.46,"pp":4195.53,"measured":true,"old_build":false,"id":"rtx-5060-ti-16","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 5060 Ti (8 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":8,"bandwidth":448,"tg":93.46,"pp":4195.53,"measured":true,"old_build":false,"id":"rtx-5060-ti-8","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 5060 (8 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":8,"bandwidth":448,"tg":93,"pp":3500,"measured":false,"old_build":false,"id":"rtx-5060","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 4090 (24 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":24,"bandwidth":1008,"tg":188.96,"pp":14770.63,"measured":true,"old_build":false,"id":"rtx-4090","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 4080 SUPER (16 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":16,"bandwidth":736,"tg":147.48,"pp":9439.01,"measured":true,"old_build":false,"id":"rtx-4080-super","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 4080 (16 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":16,"bandwidth":716.8,"tg":143.47,"pp":9205.93,"measured":true,"old_build":false,"id":"rtx-4080","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 4070 Ti SUPER (16 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":16,"bandwidth":672,"tg":132.85,"pp":7612.32,"measured":true,"old_build":false,"id":"rtx-4070-ti-super","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 4070 Ti (12 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":12,"bandwidth":504,"tg":110.53,"pp":4981.44,"measured":true,"old_build":false,"id":"rtx-4070-ti","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"GeForce RTX 4070 SUPER (12 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":12,"bandwidth":504,"tg":108.74,"pp":4608.2,"measured":true,"old_build":false,"id":"rtx-4070-super","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"GeForce RTX 4070 (12 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":12,"bandwidth":504,"tg":92.29,"pp":3179.37,"measured":true,"old_build":false,"id":"rtx-4070","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"GeForce RTX 4060 Ti (16 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":16,"bandwidth":288,"tg":64.03,"pp":3803.45,"measured":true,"old_build":false,"id":"rtx-4060-ti-16","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 4060 Ti (8 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":8,"bandwidth":288,"tg":64.03,"pp":3803.45,"measured":true,"old_build":false,"id":"rtx-4060-ti-8","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 4060 (8 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":8,"bandwidth":272,"tg":59,"pp":2700,"measured":false,"old_build":false,"id":"rtx-4060","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 3090 (24 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":24,"bandwidth":936,"tg":161.89,"pp":5560.06,"measured":true,"old_build":false,"id":"rtx-3090","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 3080 (10 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":10,"bandwidth":760,"tg":139.95,"pp":5569.56,"measured":true,"old_build":false,"id":"rtx-3080","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"GeForce RTX 3070 (8 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":8,"bandwidth":448,"tg":78.71,"pp":2113.02,"measured":true,"old_build":false,"id":"rtx-3070","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"GeForce RTX 3060 Ti (8 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":8,"bandwidth":448,"tg":86.74,"pp":2361.43,"measured":true,"old_build":false,"id":"rtx-3060-ti","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"GeForce RTX 3060 (12 GB)","kind":"nvidia","memory":[16,32,64,128],"vram":12,"bandwidth":360,"tg":76.92,"pp":2407.67,"measured":true,"old_build":false,"id":"rtx-3060","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"Radeon RX 9070 XT (16 GB)","kind":"amd","memory":[16,32,64,128],"vram":16,"bandwidth":640,"tg":137.11,"pp":5036.04,"measured":true,"old_build":false,"id":"rx-9070-xt","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 9070 (16 GB)","kind":"amd","memory":[16,32,64,128],"vram":16,"bandwidth":640,"tg":119.71,"pp":3164.1,"measured":true,"old_build":false,"id":"rx-9070","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 9060 XT (16 GB)","kind":"amd","memory":[16,32,64,128],"vram":16,"bandwidth":320,"tg":70.54,"pp":2141.67,"measured":true,"old_build":false,"id":"rx-9060-xt-16","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 9060 XT (8 GB)","kind":"amd","memory":[16,32,64,128],"vram":8,"bandwidth":320,"tg":70.54,"pp":2141.67,"measured":true,"old_build":false,"id":"rx-9060-xt-8","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 7900 XTX (24 GB)","kind":"amd","memory":[16,32,64,128],"vram":24,"bandwidth":960,"tg":182.63,"pp":3726.99,"measured":true,"old_build":false,"id":"rx-7900-xtx","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 7900 XT (20 GB)","kind":"amd","memory":[16,32,64,128],"vram":20,"bandwidth":800,"tg":123.18,"pp":2941.58,"measured":true,"old_build":false,"id":"rx-7900-xt","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 7900 GRE (16 GB)","kind":"amd","memory":[16,32,64,128],"vram":16,"bandwidth":576,"tg":116.11,"pp":2336.31,"measured":true,"old_build":false,"id":"rx-7900-gre","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 7800 XT (16 GB)","kind":"amd","memory":[16,32,64,128],"vram":16,"bandwidth":624,"tg":118.27,"pp":2017.33,"measured":true,"old_build":false,"id":"rx-7800-xt","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 7600 XT (16 GB)","kind":"amd","memory":[16,32,64,128],"vram":16,"bandwidth":288,"tg":53.02,"pp":840.85,"measured":true,"old_build":false,"id":"rx-7600-xt","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Radeon RX 7600 (8 GB)","kind":"amd","memory":[16,32,64,128],"vram":8,"bandwidth":288,"tg":58.03,"pp":1167.2,"measured":true,"old_build":false,"id":"rx-7600","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"AMD Ryzen AI Max+ 395","kind":"shared","memory":[64,128],"vram":0.75,"bandwidth":256,"tg":53.59,"pp":1288.96,"measured":true,"old_build":false,"id":"ryzen-ai-max","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"NVIDIA DGX Spark","kind":"shared","memory":[128],"vram":0.9,"bandwidth":273,"tg":56.74,"pp":3661.37,"measured":true,"old_build":false,"id":"dgx-spark","source":"https://github.com/ggml-org/llama.cpp/discussions/15013","variants":[]},{"name":"Intel Core Ultra 300 (integrated graphics)","kind":"shared","memory":[16,32,64],"vram":0.5,"bandwidth":120,"tg":24.45,"pp":1380.07,"measured":true,"old_build":false,"id":"core-ultra-300","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"Intel Core Ultra 200 (integrated graphics)","kind":"shared","memory":[16,32,64],"vram":0.5,"bandwidth":100,"tg":24.37,"pp":864.99,"measured":true,"old_build":false,"id":"core-ultra-200","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"AMD Ryzen AI 300 (integrated graphics)","kind":"shared","memory":[16,32,64],"vram":0.5,"bandwidth":100,"tg":22.41,"pp":479.07,"measured":true,"old_build":false,"id":"ryzen-ai-300","source":"https://github.com/ggml-org/llama.cpp/discussions/10879","variants":[]},{"name":"PC without a graphics card (DDR5 memory)","kind":"cpu","memory":[8,16,32,64,128],"vram":null,"bandwidth":89.6,"tg":12.489229053677327,"pp":80.64,"measured":false,"old_build":false,"id":"cpu-ddr5","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]},{"name":"PC without a graphics card (DDR4 memory)","kind":"cpu","memory":[8,16,32,64,128],"vram":null,"bandwidth":51.2,"tg":7.233496018494735,"pp":46.080000000000005,"measured":false,"old_build":false,"id":"cpu-ddr4","source":"https://github.com/ggml-org/llama.cpp/discussions/4167","variants":[]}],"models":[{"id":"qwen3.5-0.8b","name":"Qwen3.5 0.8B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":579615840,"vision_bytes":204987104,"mlx_weights_bytes":625229487,"params":0.9,"active":null,"layers":24,"context_length":262144,"kv_bytes_per_token":12288,"sliding_window":null,"state_bytes":18874368,"hybrid":true,"min_ram_gb":8,"arch":"hybrid","tier":"basic","url":"https://huggingface.co/Qwen/Qwen3.5-0.8B"},{"id":"qwen3.5-2b","name":"Qwen3.5 2B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":1396198496,"vision_bytes":668227136,"mlx_weights_bytes":1722271785,"params":2.3,"active":null,"layers":24,"context_length":262144,"kv_bytes_per_token":12288,"sliding_window":null,"state_bytes":18874368,"hybrid":true,"min_ram_gb":8,"arch":"hybrid","tier":"basic","url":"https://huggingface.co/Qwen/Qwen3.5-2B"},{"id":"qwen3.5-4b","name":"Qwen3.5 4B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":2783446880,"vision_bytes":672423488,"mlx_weights_bytes":3034300695,"params":4.7,"active":null,"layers":32,"context_length":262144,"kv_bytes_per_token":32768,"sliding_window":null,"state_bytes":50331648,"hybrid":true,"min_ram_gb":8,"arch":"hybrid","tier":"everyday","url":"https://huggingface.co/Qwen/Qwen3.5-4B"},{"id":"qwen3.5-9b","name":"Qwen3.5 9B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":5780090688,"vision_bytes":918165888,"mlx_weights_bytes":5950221072,"params":9.7,"active":null,"layers":32,"context_length":262144,"kv_bytes_per_token":32768,"sliding_window":null,"state_bytes":50331648,"hybrid":true,"min_ram_gb":16,"arch":"hybrid","tier":"solid","url":"https://huggingface.co/Qwen/Qwen3.5-9B"},{"id":"qwen3.5-27b","name":"Qwen3.5 27B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":17984872928,"vision_bytes":927607104,"mlx_weights_bytes":16054546159,"params":27.8,"active":null,"layers":64,"context_length":262144,"kv_bytes_per_token":65536,"sliding_window":null,"state_bytes":150994944,"hybrid":true,"min_ram_gb":32,"arch":"hybrid","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3.5-27B"},{"id":"qwen3.5-35b-a3b","name":"Qwen3.5 35B-A3B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":22285080384,"vision_bytes":899283552,"mlx_weights_bytes":20391679439,"params":36.0,"active":3.0,"layers":40,"context_length":262144,"kv_bytes_per_token":20480,"sliding_window":null,"state_bytes":62914560,"hybrid":true,"min_ram_gb":36,"arch":"hybrid_moe","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3.5-35B-A3B"},{"id":"qwen3.6-27b","name":"Qwen3.6 27B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":19095766304,"vision_bytes":931145984,"mlx_weights_bytes":16054541599,"params":27.8,"active":null,"layers":64,"context_length":262144,"kv_bytes_per_token":65536,"sliding_window":null,"state_bytes":150994944,"hybrid":true,"min_ram_gb":36,"arch":"hybrid","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3.6-27B"},{"id":"qwen3.6-35b-a3b-iq3","name":"Qwen3.6 35B-A3B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"3 bits","file_bytes":15761446528,"vision_bytes":899283296,"mlx_weights_bytes":null,"params":36.0,"active":3.0,"layers":40,"context_length":262144,"kv_bytes_per_token":20480,"sliding_window":null,"state_bytes":62914560,"hybrid":true,"min_ram_gb":24,"arch":"hybrid_moe","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},{"id":"qwen3.6-35b-a3b","name":"Qwen3.6 35B-A3B (4 bits)","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":20419565568,"vision_bytes":902822144,"mlx_weights_bytes":20402204271,"params":36.0,"active":3.0,"layers":40,"context_length":262144,"kv_bytes_per_token":20480,"sliding_window":null,"state_bytes":62914560,"hybrid":true,"min_ram_gb":32,"arch":"hybrid_moe","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B"},{"id":"qwen3.8-27b","name":"Qwen3.8 27B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":18973870528,"vision_bytes":931145888,"mlx_weights_bytes":16054541349,"params":27.8,"active":null,"layers":64,"context_length":262144,"kv_bytes_per_token":65536,"sliding_window":null,"state_bytes":150994944,"hybrid":true,"min_ram_gb":36,"arch":"hybrid","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3.8-27B"},{"id":"qwen3-4b-instruct-2507","name":"Qwen3 4B Instruct 2507","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":2497280736,"vision_bytes":null,"mlx_weights_bytes":2263022417,"params":4.0,"active":null,"layers":36,"context_length":262144,"kv_bytes_per_token":147456,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"everyday","url":"https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507"},{"id":"qwen3-4b-thinking-2507","name":"Qwen3 4B Thinking 2507","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":2497280736,"vision_bytes":null,"mlx_weights_bytes":2263022417,"params":4.0,"active":null,"layers":36,"context_length":262144,"kv_bytes_per_token":147456,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"everyday","url":"https://huggingface.co/Qwen/Qwen3-4B-Thinking-2507"},{"id":"qwen3-8b","name":"Qwen3 8B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":5027783488,"vision_bytes":null,"mlx_weights_bytes":4607835174,"params":8.2,"active":null,"layers":36,"context_length":40960,"kv_bytes_per_token":147456,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/Qwen/Qwen3-8B"},{"id":"qwen3-14b","name":"Qwen3 14B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":9001752960,"vision_bytes":null,"mlx_weights_bytes":7846402136,"params":14.8,"active":null,"layers":40,"context_length":40960,"kv_bytes_per_token":163840,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":24,"arch":"dense","tier":"solid","url":"https://huggingface.co/Qwen/Qwen3-14B"},{"id":"qwen3-32b","name":"Qwen3 32B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":19762149024,"vision_bytes":null,"mlx_weights_bytes":17406055438,"params":32.8,"active":null,"layers":64,"context_length":40960,"kv_bytes_per_token":262144,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":36,"arch":"dense","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3-32B"},{"id":"qwen3-30b-a3b-instruct-2507","name":"Qwen3 30B-A3B Instruct 2507","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":18632183808,"vision_bytes":null,"mlx_weights_bytes":17181071994,"params":30.5,"active":3.3,"layers":48,"context_length":262144,"kv_bytes_per_token":98304,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"moe","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Instruct-2507"},{"id":"qwen3-30b-a3b-thinking-2507","name":"Qwen3 30B-A3B Thinking 2507","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":18632183808,"vision_bytes":null,"mlx_weights_bytes":17181071029,"params":30.5,"active":3.3,"layers":48,"context_length":262144,"kv_bytes_per_token":98304,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"moe","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3-30B-A3B-Thinking-2507"},{"id":"qwen3-coder-30b-a3b","name":"Qwen3-Coder 30B-A3B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":18556689568,"vision_bytes":null,"mlx_weights_bytes":17181071994,"params":30.5,"active":3.3,"layers":48,"context_length":262144,"kv_bytes_per_token":98304,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"moe","tier":"strong","url":"https://huggingface.co/Qwen/Qwen3-Coder-30B-A3B-Instruct"},{"id":"glm-4-9b-0414","name":"GLM-4 9B","publisher":"Z.ai","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":6166574464,"vision_bytes":null,"mlx_weights_bytes":5288984411,"params":9.4,"active":null,"layers":40,"context_length":32768,"kv_bytes_per_token":40960,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/zai-org/GLM-4-9B-0414"},{"id":"glm-z1-9b-0414","name":"GLM-Z1 9B","publisher":"Z.ai","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":6166574752,"vision_bytes":null,"mlx_weights_bytes":5288984411,"params":9.4,"active":null,"layers":40,"context_length":32768,"kv_bytes_per_token":40960,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/zai-org/GLM-Z1-9B-0414"},{"id":"glm-4.6v-flash","name":"GLM-4.6V Flash","publisher":"Z.ai","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":6166577888,"vision_bytes":980055680,"mlx_weights_bytes":null,"params":10.3,"active":null,"layers":40,"context_length":131072,"kv_bytes_per_token":40960,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/zai-org/GLM-4.6V-Flash"},{"id":"glm-4.7-flash","name":"GLM-4.7 Flash","publisher":"Z.ai","license":"MIT","one_click":true,"quant":"Q4_K","file_bytes":18244193920,"vision_bytes":null,"mlx_weights_bytes":16852441819,"params":31.2,"active":3.0,"layers":47,"context_length":202752,"kv_bytes_per_token":54144,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"moe","tier":"strong","url":"https://huggingface.co/zai-org/GLM-4.7-Flash"},{"id":"glm-4-32b-0414","name":"GLM-4 32B","publisher":"Z.ai","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":19680022400,"vision_bytes":null,"mlx_weights_bytes":18320739902,"params":32.6,"active":null,"layers":61,"context_length":32768,"kv_bytes_per_token":62464,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"dense","tier":"strong","url":"https://huggingface.co/zai-org/GLM-4-32B-0414"},{"id":"deepseek-r1-0528-qwen3-8b","name":"DeepSeek-R1 8B","publisher":"DeepSeek","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":5027782816,"vision_bytes":null,"mlx_weights_bytes":4607835164,"params":8.2,"active":null,"layers":36,"context_length":131072,"kv_bytes_per_token":147456,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/deepseek-ai/DeepSeek-R1-0528-Qwen3-8B"},{"id":"deepseek-r1-distill-qwen-7b","name":"DeepSeek-R1 Distill Qwen 7B","publisher":"DeepSeek","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":4683073504,"vision_bytes":null,"mlx_weights_bytes":4284346255,"params":7.6,"active":null,"layers":28,"context_length":131072,"kv_bytes_per_token":57344,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"solid","url":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B"},{"id":"deepseek-r1-distill-llama-8b","name":"DeepSeek-R1 Distill Llama 8B","publisher":"DeepSeek","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":4920736608,"vision_bytes":null,"mlx_weights_bytes":4517489037,"params":8.0,"active":null,"layers":32,"context_length":131072,"kv_bytes_per_token":131072,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Llama-8B"},{"id":"deepseek-r1-distill-qwen-14b","name":"DeepSeek-R1 Distill Qwen 14B","publisher":"DeepSeek","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":8988110240,"vision_bytes":null,"mlx_weights_bytes":8309494233,"params":14.8,"active":null,"layers":48,"context_length":131072,"kv_bytes_per_token":196608,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":18,"arch":"dense","tier":"solid","url":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-14B"},{"id":"deepseek-r1-distill-qwen-32b","name":"DeepSeek-R1 Distill Qwen 32B","publisher":"DeepSeek","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":19851335840,"vision_bytes":null,"mlx_weights_bytes":18431478457,"params":32.8,"active":null,"layers":64,"context_length":131072,"kv_bytes_per_token":262144,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":36,"arch":"dense","tier":"strong","url":"https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"},{"id":"deepseek-coder-v2-lite","name":"DeepSeek-Coder-V2 Lite","publisher":"DeepSeek","license":"DeepSeek License","one_click":true,"quant":"Q4_K_M","file_bytes":10364416768,"vision_bytes":null,"mlx_weights_bytes":8840089384,"params":15.7,"active":2.4,"layers":27,"context_length":163840,"kv_bytes_per_token":31104,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"moe","tier":"solid","url":"https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct"},{"id":"gemma-4-e2b","name":"Gemma 4 E2B","publisher":"Google","license":"Apache 2.0","one_click":true,"quant":"Q4_0 (QAT)","file_bytes":3349516256,"vision_bytes":986833664,"mlx_weights_bytes":3550670554,"params":5.1,"active":2.3,"layers":35,"context_length":131072,"kv_bytes_per_token":6144,"sliding_window":{"bytes_per_token":12288,"window":512},"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"basic","url":"https://huggingface.co/google/gemma-4-E2B-it"},{"id":"gemma-4-e4b","name":"Gemma 4 E4B","publisher":"Google","license":"Apache 2.0","one_click":true,"quant":"Q4_0 (QAT)","file_bytes":5154941280,"vision_bytes":991552256,"mlx_weights_bytes":5146800534,"params":8.0,"active":4.5,"layers":42,"context_length":131072,"kv_bytes_per_token":16384,"sliding_window":{"bytes_per_token":40960,"window":512},"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"everyday","url":"https://huggingface.co/google/gemma-4-E4B-it"},{"id":"gemma-4-12b","name":"Gemma 4 12B","publisher":"Google","license":"Apache 2.0","one_click":true,"quant":"Q4_0 (QAT)","file_bytes":6975879296,"vision_bytes":175115616,"mlx_weights_bytes":6741039511,"params":12.0,"active":null,"layers":48,"context_length":262144,"kv_bytes_per_token":8192,"sliding_window":{"bytes_per_token":163840,"window":1024},"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/google/gemma-4-12B-it"},{"id":"gemma-4-26b-a4b","name":"Gemma 4 26B (A4B)","publisher":"Google","license":"Apache 2.0","one_click":true,"quant":"Q4_0 (QAT)","file_bytes":14439363584,"vision_bytes":1194828160,"mlx_weights_bytes":15341205776,"params":25.8,"active":3.8,"layers":30,"context_length":262144,"kv_bytes_per_token":10240,"sliding_window":{"bytes_per_token":102400,"window":1024},"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"moe","tier":"strong","url":"https://huggingface.co/google/gemma-4-26B-A4B-it"},{"id":"gemma-4-31b","name":"Gemma 4 31B","publisher":"Google","license":"Apache 2.0","one_click":true,"quant":"Q4_0 (QAT)","file_bytes":17651001568,"vision_bytes":1200726368,"mlx_weights_bytes":18412016676,"params":31.3,"active":null,"layers":60,"context_length":262144,"kv_bytes_per_token":40960,"sliding_window":{"bytes_per_token":409600,"window":1024},"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"dense","tier":"strong","url":"https://huggingface.co/google/gemma-4-31B-it"},{"id":"ministral-3-3b","name":"Ministral 3 3B","publisher":"Mistral AI","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":2147023008,"vision_bytes":841501856,"mlx_weights_bytes":2745210934,"params":3.8,"active":null,"layers":26,"context_length":262144,"kv_bytes_per_token":106496,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"everyday","url":"https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512"},{"id":"ministral-3-8b","name":"Ministral 3 8B","publisher":"Mistral AI","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":5198911904,"vision_bytes":858283168,"mlx_weights_bytes":5596684559,"params":8.9,"active":null,"layers":34,"context_length":262144,"kv_bytes_per_token":139264,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512"},{"id":"ministral-3-14b","name":"Ministral 3 14B","publisher":"Mistral AI","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":8239593024,"vision_bytes":879258784,"mlx_weights_bytes":8424582838,"params":13.9,"active":null,"layers":40,"context_length":262144,"kv_bytes_per_token":163840,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":24,"arch":"dense","tier":"solid","url":"https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512"},{"id":"ministral-3-14b-reasoning","name":"Ministral 3 14B Reasoning","publisher":"Mistral AI","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":8239591488,"vision_bytes":null,"mlx_weights_bytes":8424582838,"params":13.9,"active":null,"layers":40,"context_length":262144,"kv_bytes_per_token":163840,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/mistralai/Ministral-3-14B-Reasoning-2512"},{"id":"mistral-small-3.2","name":"Mistral Small 3.2 24B","publisher":"Mistral AI","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":14333915264,"vision_bytes":878054144,"mlx_weights_bytes":13260192731,"params":24.0,"active":null,"layers":40,"context_length":131072,"kv_bytes_per_token":163840,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"dense","tier":"strong","url":"https://huggingface.co/mistralai/Mistral-Small-3.2-24B-Instruct-2506"},{"id":"magistral-small-2509","name":"Magistral Small 1.2","publisher":"Mistral AI","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":14333912032,"vision_bytes":null,"mlx_weights_bytes":null,"params":24.0,"active":null,"layers":40,"context_length":131072,"kv_bytes_per_token":163840,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"dense","tier":"strong","url":"https://huggingface.co/mistralai/Magistral-Small-2509"},{"id":"devstral-small-2","name":"Devstral Small 2 24B","publisher":"Mistral AI","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":14334438272,"vision_bytes":null,"mlx_weights_bytes":15102831624,"params":24.0,"active":null,"layers":40,"context_length":393216,"kv_bytes_per_token":163840,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":32,"arch":"dense","tier":"strong","url":"https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512"},{"id":"llama-3.2-1b","name":"Llama 3.2 1B","publisher":"Meta","license":"Llama 3.2 Community","one_click":true,"quant":"Q4_K_M","file_bytes":807694464,"vision_bytes":null,"mlx_weights_bytes":695283921,"params":1.2,"active":null,"layers":16,"context_length":131072,"kv_bytes_per_token":32768,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"basic","url":"https://huggingface.co/meta-llama/Llama-3.2-1B-Instruct"},{"id":"llama-3.2-3b","name":"Llama 3.2 3B","publisher":"Meta","license":"Llama 3.2 Community","one_click":true,"quant":"Q4_K_M","file_bytes":2019377696,"vision_bytes":null,"mlx_weights_bytes":1807496278,"params":3.2,"active":null,"layers":28,"context_length":131072,"kv_bytes_per_token":114688,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"everyday","url":"https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct"},{"id":"llama-3.1-8b","name":"Llama 3.1 8B","publisher":"Meta","license":"Llama 3.1 Community","one_click":true,"quant":"Q4_K_M","file_bytes":4920739232,"vision_bytes":null,"mlx_weights_bytes":4517489037,"params":8.0,"active":null,"layers":32,"context_length":131072,"kv_bytes_per_token":131072,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct"},{"id":"llama-3.3-70b","name":"Llama 3.3 70B","publisher":"Meta","license":"Llama 3.3 Community","one_click":true,"quant":"Q4_K_M","file_bytes":42520398816,"vision_bytes":null,"mlx_weights_bytes":39688567605,"params":70.6,"active":null,"layers":80,"context_length":131072,"kv_bytes_per_token":327680,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":64,"arch":"dense","tier":"very_strong","url":"https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct"},{"id":"phi-4-mini","name":"Phi-4 mini","publisher":"Microsoft","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":2491874688,"vision_bytes":null,"mlx_weights_bytes":2158100796,"params":3.8,"active":null,"layers":32,"context_length":131072,"kv_bytes_per_token":131072,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"everyday","url":"https://huggingface.co/microsoft/Phi-4-mini-instruct"},{"id":"phi-4-mini-reasoning","name":"Phi-4 mini reasoning","publisher":"Microsoft","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":2491874848,"vision_bytes":null,"mlx_weights_bytes":2158101119,"params":3.8,"active":null,"layers":32,"context_length":131072,"kv_bytes_per_token":131072,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"everyday","url":"https://huggingface.co/microsoft/Phi-4-mini-reasoning"},{"id":"phi-4","name":"Phi-4 14B","publisher":"Microsoft","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":9053114816,"vision_bytes":null,"mlx_weights_bytes":8246633911,"params":14.7,"active":null,"layers":40,"context_length":16384,"kv_bytes_per_token":204800,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":18,"arch":"dense","tier":"solid","url":"https://huggingface.co/microsoft/phi-4"},{"id":"phi-4-reasoning-plus","name":"Phi-4 reasoning plus","publisher":"Microsoft","license":"MIT","one_click":true,"quant":"Q4_K_M","file_bytes":9053116416,"vision_bytes":null,"mlx_weights_bytes":9162827491,"params":14.7,"active":null,"layers":40,"context_length":32768,"kv_bytes_per_token":204800,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":18,"arch":"dense","tier":"solid","url":"https://huggingface.co/microsoft/Phi-4-reasoning-plus"},{"id":"gpt-oss-20b","name":"gpt-oss 20B","publisher":"OpenAI","license":"Apache 2.0","one_click":true,"quant":"MXFP4","file_bytes":12109566624,"vision_bytes":null,"mlx_weights_bytes":12076207568,"params":20.9,"active":3.6,"layers":24,"context_length":131072,"kv_bytes_per_token":24576,"sliding_window":{"bytes_per_token":24576,"window":128},"state_bytes":0,"hybrid":false,"min_ram_gb":24,"arch":"moe","tier":"strong","url":"https://huggingface.co/openai/gpt-oss-20b"},{"id":"nemotron-3-nano-4b","name":"Nemotron 3 Nano 4B","publisher":"NVIDIA","license":"NVIDIA Nemotron Open Model License","one_click":true,"quant":"Q4_K_M","file_bytes":2837072864,"vision_bytes":null,"mlx_weights_bytes":2237078664,"params":4.0,"active":null,"layers":42,"context_length":262144,"kv_bytes_per_token":16384,"sliding_window":null,"state_bytes":82575360,"hybrid":true,"min_ram_gb":8,"arch":"hybrid","tier":"everyday","url":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16"},{"id":"nemotron-3-nano-30b-a3b","name":"Nemotron 3 Nano 30B-A3B","publisher":"NVIDIA","license":"NVIDIA Nemotron Open Model License","one_click":true,"quant":"Q4_K_M","file_bytes":22421827488,"vision_bytes":null,"mlx_weights_bytes":17775428300,"params":31.6,"active":3.5,"layers":52,"context_length":262144,"kv_bytes_per_token":6144,"sliding_window":null,"state_bytes":48234496,"hybrid":true,"min_ram_gb":36,"arch":"hybrid_moe","tier":"strong","url":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16"},{"id":"smollm3-3b","name":"SmolLM3 3B","publisher":"Hugging Face","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":1915305312,"vision_bytes":null,"mlx_weights_bytes":1730051993,"params":3.1,"active":null,"layers":36,"context_length":65536,"kv_bytes_per_token":73728,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"everyday","url":"https://huggingface.co/HuggingFaceTB/SmolLM3-3B"},{"id":"granite-4.1-8b","name":"Granite 4.1 8B","publisher":"IBM","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":5539279904,"vision_bytes":null,"mlx_weights_bytes":5238406779,"params":8.8,"active":null,"layers":40,"context_length":131072,"kv_bytes_per_token":163840,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/ibm-granite/granite-4.1-8b"},{"id":"olmo-3-7b-instruct","name":"Olmo 3 7B","publisher":"Ai2","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":4471601792,"vision_bytes":null,"mlx_weights_bytes":4105983200,"params":7.3,"active":null,"layers":32,"context_length":65536,"kv_bytes_per_token":131072,"sliding_window":{"bytes_per_token":393216,"window":4096},"state_bytes":0,"hybrid":false,"min_ram_gb":16,"arch":"dense","tier":"solid","url":"https://huggingface.co/allenai/Olmo-3-7B-Instruct"},{"id":"lfm2.5-2.6b","name":"LFM2.5 2.6B","publisher":"Liquid AI","license":"LFM Open License v1.0","one_click":true,"quant":"Q4_K_M","file_bytes":1684023584,"vision_bytes":null,"mlx_weights_bytes":1517625850,"params":2.7,"active":null,"layers":30,"context_length":128000,"kv_bytes_per_token":16384,"sliding_window":null,"state_bytes":540672,"hybrid":true,"min_ram_gb":8,"arch":"hybrid","tier":"basic","url":"https://huggingface.co/LiquidAI/LFM2.5-2.6B"},{"id":"minicpm5-2b","name":"MiniCPM5 2B","publisher":"OpenBMB","license":"Apache 2.0","one_click":true,"quant":"Q4_K_M","file_bytes":1615826144,"vision_bytes":null,"mlx_weights_bytes":null,"params":2.5,"active":null,"layers":42,"context_length":131072,"kv_bytes_per_token":43008,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":8,"arch":"dense","tier":"basic","url":"https://huggingface.co/openbmb/MiniCPM5-2B"},{"id":"qwen3-next-80b-a3b","name":"Qwen3-Next 80B-A3B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":false,"quant":"Q4_K_M","file_bytes":48410988384,"vision_bytes":null,"mlx_weights_bytes":null,"params":81.3,"active":3.0,"layers":48,"context_length":262144,"kv_bytes_per_token":24576,"sliding_window":null,"state_bytes":75497472,"hybrid":true,"min_ram_gb":64,"arch":"hybrid_moe","tier":"very_strong","url":"https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Instruct"},{"id":"qwen3.5-122b-a10b","name":"Qwen3.5 122B-A10B","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":false,"quant":"Q4_K_M","file_bytes":77616511296,"vision_bytes":908724576,"mlx_weights_bytes":null,"params":125.1,"active":10.0,"layers":48,"context_length":262144,"kv_bytes_per_token":24576,"sliding_window":null,"state_bytes":150994944,"hybrid":true,"min_ram_gb":128,"arch":"hybrid_moe","tier":"very_strong","url":"https://huggingface.co/Qwen/Qwen3.5-122B-A10B"},{"id":"qwen3-235b-a22b-instruct-2507","name":"Qwen3 235B-A22B Instruct 2507","publisher":"Qwen (Alibaba)","license":"Apache 2.0","one_click":false,"quant":"Q4_K_M","file_bytes":142646905312,"vision_bytes":null,"mlx_weights_bytes":null,"params":235.1,"active":22.0,"layers":94,"context_length":262144,"kv_bytes_per_token":192512,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":192,"arch":"moe","tier":"frontier","url":"https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"},{"id":"gpt-oss-120b","name":"gpt-oss 120B","publisher":"OpenAI","license":"Apache 2.0","one_click":false,"quant":"MXFP4","file_bytes":63387346208,"vision_bytes":null,"mlx_weights_bytes":null,"params":116.8,"active":5.1,"layers":36,"context_length":131072,"kv_bytes_per_token":36864,"sliding_window":{"bytes_per_token":36864,"window":128},"state_bytes":0,"hybrid":false,"min_ram_gb":96,"arch":"moe","tier":"very_strong","url":"https://huggingface.co/openai/gpt-oss-120b"},{"id":"mistral-small-4-119b","name":"Mistral Small 4 119B","publisher":"Mistral AI","license":"Apache 2.0","one_click":false,"quant":"Q4_K_M","file_bytes":72636771776,"vision_bytes":857078016,"mlx_weights_bytes":null,"params":119.4,"active":6.5,"layers":36,"context_length":1048576,"kv_bytes_per_token":23040,"sliding_window":null,"state_bytes":0,"hybrid":false,"min_ram_gb":96,"arch":"moe","tier":"very_strong","url":"https://huggingface.co/mistralai/Mistral-Small-4-119B-2603"}]}