{
 "id": "kolibri-1",
 "name": "Kolibri-1",
 "repo": "Aleph-Alpha/Kolibri-1",
 "lab": {
  "name": "Aleph Alpha",
  "country": "Germany"
 },
 "released": "2026-10-02",
 "parameters": 78.1,
 "activeParameters": 3.46,
 "maxContext": 262144,
 "architecture": "MoE 384 experts · 10 full-attention layers + 40 sliding-window layers (513 tokens), GQA 4 KV × 128",
 "page": "https://studiotvai.com/vram-requirements/kolibri-1",
 "calculator": "https://studiotvai.com/llm-vram-calculator?model=kolibri-1",
 "weightsGiB": {
  "native": 73.43,
  "bf16": 145.47,
  "fp8": 73.35,
  "q4_k_m": 43.98
 },
 "kvCacheGiBPerRequest": [
  {
   "context": 8192,
   "bf16": 0.195,
   "fp8": 0.098
  },
  {
   "context": 32768,
   "bf16": 0.664,
   "fp8": 0.332
  },
  {
   "context": 131072,
   "bf16": 2.539,
   "fp8": 1.27
  },
  {
   "context": 262144,
   "bf16": 5.039,
   "fp8": 2.52
  }
 ],
 "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": {
     "native": 4,
     "bf16": 7,
     "fp8": 4,
     "q4_k_m": 3
    }
   },
   {
    "gpu": "rtx-5090",
    "name": "GeForce RTX 5090",
    "memoryGB": 32,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 3,
     "bf16": 6,
     "fp8": 3,
     "q4_k_m": 2
    }
   },
   {
    "gpu": "rtx-pro-6000",
    "name": "RTX PRO 6000 Blackwell",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 1,
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "mac-m5-max-128",
    "name": "Mac M5 Max 128 GB",
    "memoryGB": 96,
    "engine": "llama.cpp",
    "byFormat": {
     "native": 1,
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h100",
    "name": "NVIDIA H100",
    "memoryGB": 80,
    "engine": "vLLM",
    "byFormat": {
     "native": 2,
     "bf16": 3,
     "fp8": 2,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "h200",
    "name": "NVIDIA H200",
    "memoryGB": 141,
    "engine": "vLLM",
    "byFormat": {
     "native": 1,
     "bf16": 2,
     "fp8": 1,
     "q4_k_m": 1
    }
   },
   {
    "gpu": "b200",
    "name": "NVIDIA B200 (HGX)",
    "memoryGB": 180,
    "engine": "vLLM",
    "byFormat": {
     "native": 1,
     "bf16": 1,
     "fp8": 1,
     "q4_k_m": 1
    }
   }
  ]
 },
 "ggufFiles": {
  "repo": "Hob-forge/Kolibri-1-GGUF",
  "gib": {
   "Q2_K": 26.67,
   "Q3_K_M": 34.95,
   "Q4_K_M": 44.2,
   "Q5_K_M": 51.77,
   "Q6_K": 59.82,
   "Q8_0": 77.43
  }
 },
 "runsWith": [
  "vLLM",
  "MLX (Mac)"
 ],
 "verified": null,
 "updated": "2026-10-10T15:05:26.095Z"
}