{
 "schema": "local-ai-bench/result/v1",
 "run_id": "20261004T080128Z-llama.cpp-qwen3.8-flash-next-iq4xs-128k-3x-fit-fitt2048-g210-ub2048-gpu012-lab",
 "ticket": "T11",
 "kind": "lab",
 "model": "qwen3.8-flash-next",
 "config_id": "llama.cpp-qwen3.8-flash-next-iq4xs-128k-3x-fit-fitt2048-g210-ub2048-gpu012",
 "outcome": "pass",
 "failed_stage": null,
 "exit_status": 0,
 "started_at": "2026-10-04T08:01:28.586886+00:00",
 "finished_at": "2026-10-04T08:06:57.123307+00:00",
 "launch": {
  "file": "launches/llama.cpp-qwen3.8-flash-next-iq4xs-128k-3x-fit-fitt2048-g210-ub2048.json",
  "sha256": "e882737bdeeadb764f93582d803f25626630b8b0a09fb20716d71d1c2c7be516",
  "image": "ghcr.io/ggml-org/llama.cpp:server-cuda12-b11312@sha256:6cdf9529493b9581c421fec5628dcd4b8dd7c09507fc7d27ad739cc9205d66f5",
  "argv": [
   "/app/llama-server",
   "--model",
   "/models/Qwen3.8-Flash-Next-IQ4_XS/Qwen3.8-Flash-Next-IQ4_XS-00001-of-00003.gguf",
   "--ctx-size",
   "131072",
   "--flash-attn",
   "on",
   "--cache-type-k",
   "q8_0",
   "--cache-type-v",
   "q8_0",
   "--fit",
   "on",
   "--parallel",
   "1",
   "--threads",
   "15",
   "--jinja",
   "--no-mmproj",
   "--metrics",
   "--host",
   "0.0.0.0",
   "--port",
   "8080",
   "--split-mode",
   "layer",
   "--fit-target",
   "2048",
   "--batch-size",
   "2048",
   "--ubatch-size",
   "2048"
  ],
  "env": {
   "CUDA_DEVICE_ORDER": "PCI_BUS_ID",
   "CUDA_VISIBLE_DEVICES": "2,1,0"
  },
  "engine": "llama.cpp",
  "engine_source_commit": null,
  "mtp": null,
  "entrypoint": "/app/llama-server",
  "port": 8080,
  "shm": "16g",
  "flags": [],
  "weights_at": [
   "/models"
  ]
 },
 "weights": [
  {
   "repo": "bartowski/Qwen3.8-Flash-Next-GGUF",
   "revision": "928589fdb66c6ff07f22ac561e3fbce76553548f",
   "manifest": "models/bartowski-Qwen3.8-Flash-Next-IQ4_XS.json"
  }
 ],
 "host": {
  "hostname": "omarchy-gpu",
  "cpu": "AMD Ryzen 9 5950X 16-Core Processor",
  "ram_total_gib": 125.7,
  "ram_layout": "DIMM_A1 32 GiB DDR4 3200 MT/s, DIMM_A2 32 GiB DDR4 3200 MT/s, DIMM_B1 32 GiB DDR4 3200 MT/s, DIMM_B2 32 GiB DDR4 3200 MT/s (AM4: 2 channels)",
  "kernel": "7.2.5-3-omarchy",
  "driver": "610.57.04",
  "cuda": "13.3",
  "cpus_online": "0-1,3-17,19-31",
  "cpus_offline": "2,18",
  "machine_checks_this_boot": 6
 },
 "gpus": [
  {
   "index": 0,
   "uuid": "GPU-816b43e4-8d65-dfd7-2234-8517b3dfbf2d",
   "name": "NVIDIA GeForce RTX 3090",
   "bus_id": "00000000:04:00.0",
   "pcie_gen_max": 4,
   "pcie_width": 8,
   "display_active": false,
   "used_by_run": true
  },
  {
   "index": 1,
   "uuid": "GPU-c67ac872-3371-a88c-aa92-d969daf6405e",
   "name": "NVIDIA GeForce RTX 3090",
   "bus_id": "00000000:0B:00.0",
   "pcie_gen_max": 4,
   "pcie_width": 8,
   "display_active": false,
   "used_by_run": true
  },
  {
   "index": 2,
   "uuid": "GPU-76d3c6af-7f18-69d2-3145-899103de1722",
   "name": "NVIDIA GeForce RTX 3090",
   "bus_id": "00000000:0C:00.0",
   "pcie_gen_max": 4,
   "pcie_width": 8,
   "display_active": false,
   "used_by_run": true
  }
 ],
 "layout": {
  "name": "current",
  "cards": 3,
  "display": "off"
 },
 "context": {
  "configured": 131072,
  "occupied_max": {
   "value": 106294,
   "unit": "tokens"
  },
  "headroom_min": {
   "state": "unavailable",
   "reason": "no occupancy measured in this run"
  }
 },
 "cache_state": {
  "cold": true,
  "reset": [
   "container",
   "prefix_kv",
   "expert_cache",
   "ngram_row_cache",
   "page_cache"
  ]
 },
 "provenance": {
  "bench_commit": "65a204856ba6e28ca395fa6aab47a88e41c62a1a",
  "bench_dirty": true,
  "pins_sha256": "858564b6e07125ac48c2ef6c5178d0a53596f21eb5f0ebda43906472b01b4d78",
  "registry_commit": "d21258dd744e7c78be28177c90060c6af8e10b7e",
  "kit_commit": "ef883d269e50ecbea290f919f095e2f3ca633b42",
  "harness_commit": "04d809ceab9df28f9adaed044884180159172930"
 },
 "metrics": {
  "gates_passed": {
   "value": [
    "load",
    "chat",
    "reasoning",
    "tools",
    "context",
    "speed"
   ],
   "unit": "gates",
   "detail": {
    "load": true,
    "chat": true,
    "reasoning": true,
    "tools": true,
    "context": true,
    "speed": true
   }
  },
  "lab_decode_c1_tps": {
   "value": 44.3,
   "unit": "tok/s"
  },
  "lab_prefill_tps": {
   "value": 717,
   "unit": "tok/s",
   "detail": "context-gate prompt tokens / total request seconds"
  },
  "mtp_accepted_tps": {
   "state": "skipped",
   "reason": "speculative decoding off in this launch"
  },
  "mtp_acceptance_rate": {
   "state": "skipped",
   "reason": "speculative decoding off in this launch"
  }
 },
 "eligibility": "none",
 "artifacts": [
  "artifacts/runs/20261004T080128Z-lab-llama.cpp-qwen3.8-flash-next-iq4xs-128k-3x-fit-fitt2048-g210-lab"
 ],
 "notes": "served=/models/Qwen3.8-Flash-Next-IQ4_XS/Qwen3.8-Flash-Next-IQ4_XS-00001-of-00003.gguf"
}