{
  "schema_version": "sensebench-run-v2",
  "run_id": "vllm-qwen3.5-122b-a10b-fp8-h200-p002-lexen-v1-20260614",
  "created_at": "2026-06-14T21:01:03.796592+00:00",
  "git_commit": null,
  "runner": {
    "github_handle": "vassiliphilippov",
    "name": "Vassili Philippov",
    "contact": null
  },
  "dataset": {
    "dataset_id": "lexen",
    "dataset_version": "lexen-v1",
    "dataset_revision": null,
    "content_hash": "sha256:5fd4382b93f19087a1e31f6dd7d1db17c1eb17ff80fcbe1d3fdd55c0c3ecefe8",
    "item_count": 4861
  },
  "prompt": {
    "id": "p002",
    "sensebench_version": "0.1.0"
  },
  "model": {
    "kind": "self_hosted_llm",
    "display_name": "Qwen/Qwen3.5-122B-A10B",
    "requested_model": "hosted_vllm/Qwen/Qwen3.5-122B-A10B",
    "resolved_model": "Qwen/Qwen3.5-122B-A10B",
    "resolved_model_counts": {
      "Qwen/Qwen3.5-122B-A10B": 4861
    },
    "llm_vendor": "Alibaba",
    "source_kind": "open_source",
    "license": "apache-2.0",
    "model_url": "https://huggingface.co/Qwen/Qwen3.5-122B-A10B",
    "hf_revision": "a099dee70ccfcd8d5dda56aaa0b60cb8ecadabc9",
    "quantization": "fp8",
    "inference_engine": "vllm",
    "inference_engine_version": "0.22.1",
    "container_image": "vllm/vllm-openai:v0.22.1",
    "serve_command": "exec vllm serve Qwen/Qwen3.5-122B-A10B-FP8 --port 8000 --served-model-name Qwen/Qwen3.5-122B-A10B --max-model-len 8192 --gpu-memory-utilization 0.90 --language-model-only --max-num-seqs 16 --trust-remote-code --default-chat-template-kwargs \\{\\\"enable_thinking\\\":\\ false\\} --gpu-memory-utilization 0.96 --revision a099dee70ccfcd8d5dda56aaa0b60cb8ecadabc9",
    "endpoint_base_url": "http://localhost:8000/v1"
  },
  "sampling": {
    "temperature": 0.0,
    "top_p": null,
    "max_tokens": 128,
    "seed": null,
    "extra": {}
  },
  "policy": {
    "votes_per_item": 1,
    "semantic_reasks_per_invalid_vote": 1,
    "tie_break": "earliest_vote",
    "monosemous_policy": "short_circuit"
  },
  "machine": {
    "platform": "Linux-6.8.0-107-generic-x86_64-with-glibc2.35",
    "cpu_model": "Intel(R) Xeon(R) Platinum 8480+",
    "cpu_cores": 224,
    "ram_gib": 2015.3,
    "gpu": {
      "name": "NVIDIA H200",
      "count": 1,
      "vram_mib_per_gpu": 143771,
      "driver_version": "595.71.05",
      "cuda_version": "13.2"
    },
    "provider": "vast.ai",
    "instance_id": "40919270",
    "hourly_rate_usd": 3.824451754385964
  },
  "execution": {
    "concurrency": 64,
    "warmup_call_count": 8,
    "timing": {
      "benchmark_started_at": "2026-06-14T20:59:33.907572+00:00",
      "benchmark_ended_at": "2026-06-14T21:01:03.790201+00:00",
      "benchmark_seconds": 89.88262075511739,
      "setup_seconds": 2.3348206968512386
    }
  },
  "totals": {
    "item_count": 4861,
    "correct_count": 4155,
    "accuracy": 0.8547623945690187,
    "call_count": 4861,
    "usage": {
      "input_tokens": 1176454,
      "cached_input_tokens": null,
      "output_tokens": 9886,
      "reasoning_output_tokens": null
    },
    "cost": {
      "currency": "USD",
      "total_usd": 0.09548659628769916,
      "input_uncached_usd": null,
      "input_cached_usd": null,
      "output_usd": null,
      "input_uncached_unit_price_usd": null,
      "input_cached_unit_price_usd": null,
      "output_unit_price_usd": null,
      "source": "machine_time_estimate"
    },
    "elapsed_seconds": 89.88262075511739
  }
}