{
 "schema": "local-ai-bench/result/v1",
 "run_id": "20261008T061821Z-strata-qwen3.8-flash-next-gsq-rco-q2_0-128k-rtx-3090-24gb-gpu2-stability",
 "ticket": "T41",
 "kind": "stability",
 "model": "qwen3.8-flash-next",
 "config_id": "strata-qwen3.8-flash-next-gsq-rco-q2_0-128k-rtx-3090-24gb-gpu2",
 "outcome": "pass",
 "failed_stage": null,
 "exit_status": 0,
 "started_at": "2026-10-08T06:18:21.959324+00:00",
 "finished_at": "2026-10-08T06:29:00.371813+00:00",
 "launch": {
  "file": "launches/strata-qwen3.8-flash-next-gsq-rco-q2_0-128k-rtx-3090-24gb.json",
  "sha256": "aadd2d65b0c6365622705da0d70416e9f7ede7673d2676c9a02bb8ae37403e0f",
  "image": "sha256:dcc2e2e0893d790f843251b5031a8070f5bc4db2389218fbe52f4b484e9ce172",
  "argv": [],
  "env": {
   "FAMILY": "qwen",
   "MODEL": "Q2_0",
   "GGUF_DIR": "/gguf/Q2_0",
   "CONTEXT": "131072",
   "VISION": "no",
   "KV": "int8",
   "GPU": "0",
   "LOW_RAM": "off",
   "PORT": "8080",
   "CUDA_DEVICE_ORDER": "PCI_BUS_ID"
  },
  "engine": "strata",
  "engine_source_commit": "https://github.com/Niko1221/Strata/tree/d5ea7133741e67743c0e886bb426c0ce8d69cf6c",
  "mtp": null,
  "entrypoint": null,
  "port": 8080,
  "shm": "16g",
  "flags": [
   "--ulimit",
   "memlock=-1",
   "--stop-timeout",
   "60"
  ],
  "weights_at": [
   "/gguf"
  ]
 },
 "weights": [
  {
   "repo": "ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF",
   "revision": "ed59f92082b1e93c0e96d60a8b11aab089b52f09",
   "manifest": "models/ISTA-DASLab-Qwen3.8-Flash-Next-GSQ-RCO-GGUF.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",
  "cpu_boost": false,
  "cpu_max_mhz": 3401,
  "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": false
  },
  {
   "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": true,
   "used_by_run": false
  },
  {
   "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": 1,
  "display": "on"
 },
 "context": {
  "configured": 131072,
  "occupied_max": {
   "value": 38712,
   "unit": "tokens",
   "detail": {
    "source": "max usage.prompt_tokens over the run"
   }
  },
  "headroom_min": {
   "value": 92079,
   "unit": "tokens",
   "detail": {
    "configured": 131072,
    "configured_source": "env CONFIGURED_CTX",
    "max_prompt_plus_completion": 38993
   }
  }
 },
 "cache_state": {
  "cold": false,
  "reset": []
 },
 "provenance": {
  "bench_commit": "8050e8e736925db26a1dcb695f1a33034f8e7d84",
  "bench_dirty": true,
  "pins_sha256": "858564b6e07125ac48c2ef6c5178d0a53596f21eb5f0ebda43906472b01b4d78",
  "registry_commit": "d21258dd744e7c78be28177c90060c6af8e10b7e",
  "kit_commit": "ef883d269e50ecbea290f919f095e2f3ca633b42",
  "harness_commit": "04d809ceab9df28f9adaed044884180159172930"
 },
 "metrics": {
  "duration_s": {
   "value": 600.0,
   "unit": "s",
   "detail": {
    "cap_s": 600.0,
    "driver_capped": true
   }
  },
  "tool_calls": {
   "value": 116,
   "unit": "calls",
   "detail": {
    "per_min": 11.6,
    "requests": 97
   }
  },
  "parse_failures": {
   "value": 0,
   "unit": "responses",
   "detail": {
    "wire": {},
    "harness_format_errors": {},
    "union_rule": "a response counts once even if both the wire check and the harness flag it",
    "counted": "malformed tool calls only (decision 0003); truncations and other format errors are separate metrics"
   }
  },
  "output_truncations": {
   "value": 0,
   "unit": "responses",
   "detail": {
    "rule": "cut off at max_tokens with no tool call"
   }
  },
  "agent_format_errors": {
   "value": 0,
   "unit": "responses",
   "detail": {
    "rule": "agent protocol errors other than malformed tool calls"
   }
  },
  "repetition_hits": {
   "value": 0,
   "unit": "hits",
   "detail": {
    "ngram64x3": 0,
    "same_call_x4": 0,
    "ngram64x3_cross_field_not_counted": 0,
    "tokenizer": "tiktoken cl100k_base (approximation: server exposes no tokenize endpoint)",
    "saved": "artifacts: proxy/repetition/"
   }
  },
  "decode_window_min_tps": {
   "value": 123.348,
   "unit": "tok/s",
   "detail": {
    "window_s": 60,
    "skip_s": 60,
    "rolling_step_s": 1,
    "n_rolling_windows": 151,
    "min_at_generation_s": 111.0,
    "windows_tps": [
     128.04,
     124.61,
     127.52
    ],
    "windows_tps_note": "consecutive non-overlapping 60 s windows; the value is the minimum over rolling windows (1 s step)"
   }
  },
  "decode_whole_run_tps": {
   "value": 126.656,
   "unit": "tok/s",
   "detail": {
    "generation_seconds": 270.449,
    "generated_tokens": 34254.1
   }
  },
  "tasks_attempted": {
   "value": 3,
   "unit": "attempts",
   "detail": {
    "unfinished_at_cap": 1,
    "task_ids": [
     "django__django-11099",
     "sympy__sympy-21612",
     "psf__requests-2317",
     "matplotlib__matplotlib-23299",
     "scikit-learn__scikit-learn-13439",
     "astropy__astropy-12907",
     "pytest-dev__pytest-7373",
     "sympy__sympy-20590"
    ],
    "task_set_sha256": "8e7487ca7d059df6359fc7d9b86c7021256a8a81b8af049616750f8f83400324"
   }
  },
  "tasks_solved": {
   "value": 2,
   "unit": "attempts",
   "detail": {
    "eval_errors": 0,
    "swebench": "5.0.2"
   }
  },
  "crash_or_oom": {
   "value": false,
   "unit": "bool",
   "detail": {
    "upstream_connection_failures": 0,
    "upstream_http_errors": 0,
    "source": "proxy-observed; the run wrapper adds container death/OOM"
   }
  },
  "gates_after": {
   "state": "skipped",
   "reason": "filled by qualify step: the lab run that follows this run (tools/qualify.py)"
  },
  "agent_attempts": {
   "value": [
    {
     "attempt": 1,
     "round": 1,
     "instance_id": "django__django-11099",
     "exit_status": "Submitted",
     "submitted": true,
     "resolved": true,
     "seconds": 52.3,
     "requests": 18,
     "tool_calls": 28,
     "completion_tokens": 3181,
     "occupied_max": 12800
    },
    {
     "attempt": 2,
     "round": 1,
     "instance_id": "sympy__sympy-21612",
     "exit_status": "Submitted",
     "submitted": true,
     "resolved": true,
     "seconds": 240.9,
     "requests": 49,
     "tool_calls": 57,
     "completion_tokens": 18657,
     "occupied_max": 38712
    },
    {
     "attempt": 3,
     "round": 1,
     "instance_id": "psf__requests-2317",
     "exit_status": "WallClockCap",
     "submitted": false,
     "resolved": false,
     "seconds": 307.1,
     "requests": 30,
     "tool_calls": 31,
     "completion_tokens": 12531,
     "occupied_max": 27651
    }
   ],
   "unit": "list"
  },
  "tasks_over_64k": {
   "value": [],
   "unit": "instance ids"
  },
  "completion_tokens": {
   "value": 34369,
   "unit": "tokens",
   "detail": {
    "finish_reasons": {
     "tool_calls": 97
    }
   }
  },
  "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/20261008T061821Z-lab-strata-qwen3.8-flash-next-gsq-rco-q2_0-128k-rtx-3090-24gb-gp-stability"
 ],
 "notes": "stability run, 10 min cap, floor 50.0 tok/s; mini-swe-agent 04d809ceab9d (benchmarks/swebench.yaml, native tool calls); sampling {\"temperature\": 0.7, \"top_p\": 0.95, \"seed\": 1234, \"max_tokens\": 8192} | repetition tokenizer: tiktoken cl100k_base (approximation: server exposes no tokenize endpoint) | 116 tool calls in this run count toward the config's 100-call minimum (\u00a75)"
}
