{
 "id": "mamba-codestral-7b",
 "name": "Mamba-Codestral 7B",
 "repo": "mistralai/Mamba-Codestral-7B-v0.1",
 "lab": {
  "name": "Mistral AI",
  "country": "France"
 },
 "released": "2024-07-16",
 "parameters": 7.29,
 "activeParameters": null,
 "maxContext": 262144,
 "architecture": "64 Mamba-2 layers, no attention: fixed state instead of a KV cache",
 "page": "https://studiotvai.com/vram-requirements/mamba-codestral-7b",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=mamba-codestral-7b",
 "weightsGiB": {
  "bf16": 13.58,
  "fp8": 7.04,
  "q4_k_m": 4.13
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 0,
   "fp8": 0
  },
  {
   "context": 32768,
   "bf16": 0,
   "fp8": 0
  },
  {
   "context": 131072,
   "bf16": 0,
   "fp8": 0
  },
  {
   "context": 262144,
   "bf16": 0,
   "fp8": 0
  }
 ],
 "recurrentStateGiBPerRequest": 0.2537,
 "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": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 1,
     "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": "strix-halo",
    "name": "Ryzen AI Max+ 395 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "dgx-spark",
    "name": "DGX Spark (GB10)",
    "memoryGB": 128,
    "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": "gabriellarson/Mamba-Codestral-7B-v0.1-GGUF",
  "gib": {
   "F16": 13.58,
   "IQ1_M": 1.57,
   "IQ1_S": 1.42,
   "IQ2_M": 2.25,
   "IQ2_S": 2.04,
   "IQ2_XS": 2.03,
   "IQ2_XXS": 1.83,
   "IQ3_M": 2.98,
   "IQ3_XXS": 2.66,
   "IQ4_NL": 3.86,
   "IQ4_XS": 3.65,
   "Q2_K": 2.3,
   "Q3_K_M": 2.98,
   "Q4_0": 3.86,
   "Q5_0": 4.69,
   "Q6_K": 5.58,
   "Q8_0": 7.22
  }
 },
 "runsWith": [],
 "verified": null,
 "updated": "2026-10-30T00:07:05.915Z"
}