{
 "id": "qwen3-32b",
 "name": "Qwen3 32B",
 "repo": "Qwen/Qwen3-32B",
 "lab": {
  "name": "Alibaba",
  "country": "China"
 },
 "released": "2025-04-27",
 "parameters": 32.76,
 "activeParameters": null,
 "maxContext": 131072,
 "architecture": "64 layers · GQA, 8 KV heads × 128 · 32K native, 128K with YaRN",
 "page": "https://studiotvai.com/vram-requirements/qwen3-32b",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=qwen3-32b",
 "weightsGiB": {
  "bf16": 61.02,
  "fp8": 31.96,
  "q4_k_m": 18.58
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 2,
   "fp8": 1
  },
  {
   "context": 32768,
   "bf16": 8,
   "fp8": 4
  },
  {
   "context": 131072,
   "bf16": 32,
   "fp8": 16
  }
 ],
 "recurrentStateGiBPerRequest": 0,
 "gpusNeeded": {
  "scenario": "one request of 8192 tokens, BF16 KV cache; vLLM on data-center GPUs, llama.cpp elsewhere",
  "gpus": [
   {
    "gpu": "rtx-4090",
    "name": "GeForce RTX 4090",
    "memoryGB": 24,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 3,
     "fp8": 2,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 3,
     "fp8": 2,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "mac-m5-max-128",
    "name": "Mac M5 Max 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "b200",
    "name": "NVIDIA B200 (HGX)",
    "memoryGB": 180,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   }
  ]
 },
 "ggufFiles": {
  "repo": "Qwen/Qwen3-32B-GGUF",
  "gib": {
   "Q4_K_M": 18.4,
   "Q5_0": 21.08,
   "Q5_K_M": 21.62,
   "Q6_K": 25.04,
   "Q8_0": 32.43
  }
 },
 "runsWith": [
  "vLLM",
  "SGLang",
  "llama.cpp (Ollama, LM Studio)",
  "MLX (Mac)"
 ],
 "verified": {
  "gguf": {
   "day": "2026-10-10",
   "repo": "Qwen/Qwen3-32B-GGUF",
   "kv": "match",
   "ours": 262144,
   "theirs": 262144
  }
 },
 "updated": "2026-10-10T15:05:26.095Z"
}