{ "backends": [ { "id": "executorch-qnn-android", "platforms": [ "android" ], "selector": "executorch-qnn-android" }, { "id": "ik_llama.cpp-cuda", "platforms": [ "windows" ], "selector": "ik_llama.cpp-cuda" }, { "id": "ik_llama.cpp-linux-cuda", "platforms": [ "linux" ], "selector": "ik_llama.cpp-cuda" }, { "id": "litert-android-cpu", "platforms": [ "android" ], "selector": "litert-lm-cpu" }, { "id": "litert-android-gpu", "platforms": [ "android" ], "selector": "litert-lm-gpu" }, { "id": "llama.cpp-android", "platforms": [ "android" ], "selector": "llama.cpp-android" }, { "id": "llama.cpp-android-cpu", "platforms": [ "android" ], "selector": "llama.cpp-cpu" }, { "id": "llama.cpp-android-htp", "platforms": [ "android" ], "selector": "llama.cpp-htp" }, { "id": "llama.cpp-cpu", "platforms": [ "windows" ], "selector": "llama.cpp-cpu" }, { "id": "llama.cpp-cuda", "platforms": [ "windows" ], "selector": "llama.cpp-cuda" }, { "id": "llama.cpp-hip", "platforms": [ "windows" ], "selector": "llama.cpp-hip" }, { "id": "llama.cpp-ios", "platforms": [ "ios" ], "selector": "llama.cpp-metal" }, { "id": "llama.cpp-linux", "platforms": [ "linux" ], "selector": "llama.cpp-cpu" }, { "id": "llama.cpp-linux-cuda", "platforms": [ "linux" ], "selector": "llama.cpp-cuda" }, { "id": "llama.cpp-linux-openvino", "platforms": [ "linux" ], "selector": "llama.cpp-openvino" }, { "id": "llama.cpp-linux-rocm", "platforms": [ "linux" ], "selector": "llama.cpp-rocm" }, { "id": "llama.cpp-linux-vulkan", "platforms": [ "linux" ], "selector": "llama.cpp-vulkan" }, { "id": "llama.cpp-mac", "platforms": [ "macos" ], "selector": "llama.cpp-metal" }, { "id": "llama.cpp-mac-intel", "platforms": [ "macos" ], "selector": "llama.cpp-cpu" }, { "id": "llama.cpp-sycl", "platforms": [ "windows" ], "selector": "llama.cpp-sycl" }, { "id": "llama.cpp-vulkan", "platforms": [ "windows" ], "selector": "llama.cpp-vulkan" }, { "id": "llama.cpp-windows-arm64", "platforms": [ "windows" ], "selector": "llama.cpp-cpu-arm64" }, { "id": "mlx-ios", "platforms": [ "ios" ], "selector": "mlx-metal" }, { "id": "mlx-mac", "platforms": [ "macos" ], "selector": "mlx-metal" }, { "id": "mnn-android-cpu", "platforms": [ "android" ], "selector": "mnn-cpu" }, { "id": "mnn-android-opencl", "platforms": [ "android" ], "selector": "mnn-opencl" }, { "id": "mnn-linux", "platforms": [ "linux" ], "selector": "mnn-cpu" }, { "id": "omnicore-android", "platforms": [ "android" ], "selector": "omnicore-android" }, { "id": "omnicore-linux-cuda", "platforms": [ "linux" ], "selector": "omnicore-linux-cuda" }, { "id": "omnicore-vulkan", "platforms": [ "windows" ], "selector": "omnicore-vulkan" }, { "id": "turboquant-mac", "platforms": [ "macos" ], "selector": "turboquant-metal" }, { "id": "vllm-linux-cuda", "platforms": [ "linux" ], "selector": "vllm-cuda" }, { "id": "vllm-wsl2-cuda", "platforms": [ "windows" ], "selector": "vllm-wsl2-cuda" }, { "id": "vllm-wsl2-rocm", "platforms": [ "windows" ], "selector": "vllm-wsl2-rocm" } ], "catalog_schema_version": 1, "contract_version": "2.0.0", "data_kind": "benchmark-catalog-index", "devices": [ { "backend_ids": [ "llama.cpp-hip", "llama.cpp-linux-rocm", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-wsl2-rocm" ], "id": "amd-radeon-pc", "name": "AMD Radeon GPU", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "llama.cpp-mac", "mlx-mac", "turboquant-mac" ], "id": "apple-silicon", "name": "Apple Silicon", "platforms": [ "macos" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "nvidia-laptop-cuda", "name": "NVIDIA Laptop GPU", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "llama.cpp-hip", "llama.cpp-vulkan", "omnicore-vulkan", "vllm-wsl2-rocm" ], "id": "radeon-8060s", "name": "AMD Radeon 8060S", "platforms": [ "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-3060", "name": "NVIDIA GeForce RTX 3060", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-3060-ti", "name": "NVIDIA GeForce RTX 3060 Ti", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-3070", "name": "NVIDIA GeForce RTX 3070", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-3070-ti", "name": "NVIDIA GeForce RTX 3070 Ti", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-3080", "name": "NVIDIA GeForce RTX 3080", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-3080-ti", "name": "NVIDIA GeForce RTX 3080 Ti", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "omnicore-linux-cuda", "vllm-linux-cuda" ], "id": "rtx-3090", "name": "NVIDIA GeForce RTX 3090", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-3090-ti", "name": "NVIDIA GeForce RTX 3090 Ti", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-4060", "name": "NVIDIA GeForce RTX 4060", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-4060-ti", "name": "NVIDIA GeForce RTX 4060 Ti", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-4070", "name": "NVIDIA GeForce RTX 4070", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-4070-super", "name": "NVIDIA GeForce RTX 4070 SUPER", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-4070-ti", "name": "NVIDIA GeForce RTX 4070 Ti", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-4070-ti-super", "name": "NVIDIA GeForce RTX 4070 Ti SUPER", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-4080", "name": "NVIDIA GeForce RTX 4080", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-4080-super", "name": "NVIDIA GeForce RTX 4080 SUPER", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "omnicore-linux-cuda", "vllm-linux-cuda" ], "id": "rtx-4090", "name": "NVIDIA GeForce RTX 4090", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-5060", "name": "NVIDIA GeForce RTX 5060", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-5060-ti", "name": "NVIDIA GeForce RTX 5060 Ti", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-5070", "name": "NVIDIA GeForce RTX 5070", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-5070-ti", "name": "NVIDIA GeForce RTX 5070 Ti", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "vllm-linux-cuda" ], "id": "rtx-5080", "name": "NVIDIA GeForce RTX 5080", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "ik_llama.cpp-cuda", "ik_llama.cpp-linux-cuda", "llama.cpp-cuda", "llama.cpp-linux-cuda", "llama.cpp-linux-vulkan", "llama.cpp-vulkan", "omnicore-linux-cuda", "vllm-linux-cuda" ], "id": "rtx-5090", "name": "NVIDIA GeForce RTX 5090", "platforms": [ "linux", "windows" ] }, { "backend_ids": [ "executorch-qnn-android", "litert-android-cpu", "litert-android-gpu", "llama.cpp-android", "llama.cpp-android-cpu", "mnn-android-cpu", "mnn-android-opencl", "omnicore-android" ], "id": "sm8650", "name": "Snapdragon 8 Gen 3 · SM8650", "platforms": [ "android" ] }, { "backend_ids": [ "executorch-qnn-android", "litert-android-cpu", "litert-android-gpu", "llama.cpp-android", "llama.cpp-android-cpu", "llama.cpp-android-htp", "mnn-android-cpu", "mnn-android-opencl", "omnicore-android" ], "id": "sm8750", "name": "Snapdragon 8 Elite · SM8750", "platforms": [ "android" ] }, { "backend_ids": [ "litert-android-gpu", "llama.cpp-android", "llama.cpp-android-cpu", "llama.cpp-android-htp", "mnn-android-cpu", "mnn-android-opencl", "omnicore-android" ], "id": "sm8850", "name": "Snapdragon 8 Elite Gen 5 · SM8850", "platforms": [ "android" ] }, { "backend_ids": [ "llama.cpp-cpu", "llama.cpp-linux" ], "id": "x86-64-cpu", "name": "x86-64 PC CPU", "platforms": [ "linux", "windows" ] } ], "models": [ { "id": "deepseek-r1-0528-qwen3-8b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "deepseek-r1-distill-qwen-1-5b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "deepseek-r1-distill-qwen-14b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "deepseek-r1-distill-qwen-32b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "deepseek-r1-distill-qwen-7b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "deepseek-v4-flash", "modalities": [ "text" ], "support": [ { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "UD-IQ1_S" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "UD-IQ1_M", "UD-IQ1_S", "UD-IQ2_M", "UD-IQ2_XXS", "UD-IQ3_S", "UD-IQ3_XXS", "UD-IQ4_NL", "UD-IQ4_XS", "UD-Q2_K_XL", "UD-Q3_K_M", "UD-Q3_K_XL", "UD-Q4_K_XL", "UD-Q8_K_XL" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "FP4" ] } ] }, { "id": "gemma-4-12b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "gemma-4-26b-a4b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "gemma-4-31b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "gemma-4-e2b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "litert-android-gpu", "format": "LITERTLM", "platform_id": "android", "quantizations": [ "int4" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "gemma-4-e4b", "modalities": [ "audio", "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "glm-4-9b-0414", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "glm-5.3-flash", "modalities": [ "image", "text" ], "support": [ { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "UD-IQ1_S" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q8_0", "UD-IQ1_M", "UD-IQ1_S", "UD-IQ2_XXS", "UD-IQ3_XXS", "UD-IQ4_XS", "UD-Q2_K_XL", "UD-Q3_K_XL", "UD-Q4_K_XL", "UD-Q5_K_XL", "UD-Q6_K_XL" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "FP8" ] } ] }, { "id": "gpt-oss-20b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "minicpm5-2b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "F16", "Q4_K_M", "Q8_0" ] }, { "backend_id": "omnicore-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "F16", "Q4_K_M", "Q8_0" ] } ] }, { "id": "qwen2-5-0-5b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen2-5-1-5b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen2-5-14b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen2-5-32b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen2-5-3b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen2-5-7b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen2-5-vl-32b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q5_K_M", "Q6_K" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q5_K_M", "Q6_K" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q5_K_M", "Q6_K" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q5_K_M", "Q6_K" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q5_K_M", "Q6_K" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q5_K_M", "Q6_K" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q5_K_M", "Q6_K" ] } ] }, { "id": "qwen2-5-vl-3b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen2-5-vl-7b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-0-6b", "modalities": [ "text" ], "support": [ { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-14b", "modalities": [ "text" ], "support": [ { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-32b", "modalities": [ "text" ], "support": [ { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-4b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "omnicore-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-5-0-8b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "omnicore-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-5-122b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-5-27b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "omnicore-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-5-2b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "omnicore-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-5-35b-a3b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "omnicore-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-vulkan", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M" ] }, { "backend_id": "omnicore-vulkan", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M" ] } ] }, { "id": "qwen3-5-4b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "omnicore-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "omnicore-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-5-9b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "omnicore-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "omnicore-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-6-27b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] } ] }, { "id": "qwen3-6-35b-a3b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "omnicore-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-vulkan", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M" ] }, { "backend_id": "omnicore-vulkan", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M" ] } ] }, { "id": "qwen3-8-27b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] } ] }, { "id": "qwen3-8b", "modalities": [ "text" ], "support": [ { "backend_id": "llama.cpp-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "omnicore-android", "format": "GGUF", "platform_id": "android", "quantizations": [ "Q4_0" ] }, { "backend_id": "ik_llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "omnicore-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "Q4_K_M" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "ik_llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-vl-2b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-vl-32b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-vl-4b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] }, { "id": "qwen3-vl-8b", "modalities": [ "image", "text" ], "support": [ { "backend_id": "llama.cpp-linux", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-cuda", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-linux-vulkan", "format": "GGUF", "platform_id": "linux", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "vllm-linux-cuda", "format": "Safetensors", "platform_id": "linux", "quantizations": [ "BF16" ] }, { "backend_id": "llama.cpp-mac", "format": "GGUF", "platform_id": "macos", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "mlx-mac", "format": "MLX", "platform_id": "macos", "quantizations": [ "4bit", "8bit" ] }, { "backend_id": "llama.cpp-cpu", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-cuda", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] }, { "backend_id": "llama.cpp-hip", "format": "GGUF", "platform_id": "windows", "quantizations": [ "BF16", "Q4_K_M", "Q6_K", "Q8_0" ] } ] } ], "platforms": [ { "id": "android" }, { "id": "ios" }, { "id": "linux" }, { "id": "macos" }, { "id": "windows" } ], "schema_version": "1.5.0", "source_commit": "449582c2e0a7e6ce8cb38a73e9ff1e2547867a34", "source_timestamp": "2026-08-26T19:03:54+08:00" }