{
 "id": "qwen3-coder-30b-a3b",
 "name": "Qwen3-Coder 30B-A3B",
 "repo": "Qwen/Qwen3-Coder-30B-A3B-Instruct",
 "lab": {
  "name": "Alibaba",
  "country": "China"
 },
 "released": "2025-07-31",
 "parameters": 30.53,
 "activeParameters": 3.3,
 "maxContext": 262144,
 "architecture": "MoE 128 experts · 48 layers, full attention (GQA, 4 KV heads × 128)",
 "page": "https://studiotvai.com/vram-requirements/qwen3-coder-30b-a3b",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=qwen3-coder-30b-a3b",
 "weightsGiB": {
  "bf16": 56.87,
  "fp8": 29.01,
  "q4_k_m": 17.23
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 0.75,
   "fp8": 0.375
  },
  {
   "context": 32768,
   "bf16": 3,
   "fp8": 1.5
  },
  {
   "context": 131072,
   "bf16": 12,
   "fp8": 6
  },
  {
   "context": 262144,
   "bf16": 24,
   "fp8": 12
  }
 ],
 "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": 2,
     "fp8": 1,
     "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": "unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF",
  "gib": {
   "BF16": 56.9,
   "IQ4_NL": 16.12,
   "IQ4_XS": 15.25,
   "Q2_K": 10.49,
   "Q2_K_L": 10.55,
   "Q3_K_M": 13.7,
   "Q3_K_S": 12.38,
   "Q4_0": 16.19,
   "Q4_1": 17.87,
   "Q4_K_M": 17.28,
   "Q4_K_S": 16.26,
   "Q5_K_M": 20.23,
   "Q5_K_S": 19.63,
   "Q6_K": 23.37,
   "Q8_0": 30.25,
   "UD-IQ1_M": 8.97,
   "UD-IQ1_S": 8.3,
   "UD-IQ2_M": 10.09,
   "UD-IQ2_XXS": 9.62,
   "UD-IQ3_XXS": 11.97,
   "UD-Q2_K_XL": 10.98,
   "UD-Q3_K_XL": 12.86,
   "UD-Q4_K_XL": 16.45,
   "UD-Q5_K_XL": 20.25,
   "UD-Q6_K_XL": 24.53,
   "UD-Q8_K_XL": 33.52,
   "UD-TQ1_0": 7.46
  }
 },
 "runsWith": [
  "vLLM",
  "SGLang",
  "llama.cpp (Ollama, LM Studio)",
  "MLX (Mac)"
 ],
 "verified": {
  "gguf": {
   "day": "2026-10-10",
   "repo": "unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF",
   "kv": "match",
   "ours": 98304,
   "theirs": 98304
  }
 },
 "updated": "2026-10-10T15:05:26.095Z"
}