{
 "id": "qwen3-coder-480b-a35b",
 "name": "Qwen3-Coder 480B-A35B",
 "repo": "Qwen/Qwen3-Coder-480B-A35B-Instruct",
 "lab": {
  "name": "Alibaba",
  "country": "China"
 },
 "released": "2025-07-22",
 "parameters": 480.15,
 "activeParameters": 35,
 "maxContext": 262144,
 "architecture": "MoE 160 experts · 62 layers, full attention (GQA, 8 KV heads × 128)",
 "page": "https://studiotvai.com/vram-requirements/qwen3-coder-480b-a35b",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=qwen3-coder-480b-a35b",
 "weightsGiB": {
  "bf16": 894.35,
  "fp8": 448.91,
  "q4_k_m": 270.17
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 1.938,
   "fp8": 0.969
  },
  {
   "context": 32768,
   "bf16": 7.75,
   "fp8": 3.875
  },
  {
   "context": 131072,
   "bf16": 31,
   "fp8": 15.5
  },
  {
   "context": 262144,
   "bf16": 62,
   "fp8": 31
  }
 ],
 "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": 48,
     "fp8": 24,
     "q4_k_m": 16
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 48,
     "fp8": 24,
     "q4_k_m": 16
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 16,
     "fp8": 6,
     "q4_k_m": 4
    }
   },
   {
    "gpu": "mac-m5-max-128",
    "name": "Mac M5 Max 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 16,
     "fp8": 6,
     "q4_k_m": 4
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 16,
     "fp8": 7,
     "q4_k_m": 4
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 8,
     "fp8": 4,
     "q4_k_m": 3
    }
   },
   {
    "gpu": "b200",
    "name": "NVIDIA B200 (HGX)",
    "memoryGB": 180,
    "engine": "vLLM",
    "byFormat": {
     "bf16": 6,
     "fp8": 3,
     "q4_k_m": 2
    }
   }
  ]
 },
 "ggufFiles": {
  "repo": "unsloth/Qwen3-Coder-480B-A35B-Instruct-GGUF",
  "gib": {
   "BF16": 894.48,
   "IQ4_NL": 252.01,
   "IQ4_XS": 242.65,
   "Q2_K": 162.66,
   "Q2_K_L": 162.87,
   "Q3_K_M": 213.5,
   "Q3_K_S": 192.7,
   "Q4_0": 252.99,
   "Q4_1": 279.86,
   "Q4_K_M": 270.14,
   "Q4_K_S": 254.02,
   "Q5_K_M": 317.11,
   "Q5_K_S": 307.75,
   "Q6_K": 367.01,
   "Q8_0": 475.3,
   "UD-IQ1_M": 139.43,
   "UD-IQ3_XXS": 187.7,
   "UD-Q2_K_XL": 167.91,
   "UD-Q3_K_XL": 198.06,
   "UD-Q4_K_XL": 256.63,
   "UD-Q5_K_XL": 316.64,
   "UD-Q6_K_XL": 380.02,
   "UD-Q8_K_XL": 510.59
  }
 },
 "runsWith": [
  "vLLM",
  "SGLang",
  "llama.cpp (Ollama, LM Studio)",
  "MLX (Mac)"
 ],
 "verified": {
  "gguf": {
   "day": "2026-10-10",
   "repo": "unsloth/Qwen3-Coder-480B-A35B-Instruct-GGUF",
   "kv": "match",
   "ours": 253952,
   "theirs": 253952
  }
 },
 "updated": "2026-10-10T15:05:26.095Z"
}