{
  "schema": "computearena-public-benchmark/1",
  "submission_id": "081704d1-4470-4c91-b2b9-f68c97bc9cd6",
  "run_id": "b38c68e519f4f7da5652d47770682dd3",
  "created_at_unix_ms": 1789839064595,
  "submitted_by": "prabod",
  "model": {
    "identity_resolution": "unresolved",
    "identity_verification": "unresolved",
    "name": "Llama-3.2-3B-Instruct-cuda-q8",
    "architecture": "llama",
    "quantization": "base_q4",
    "artifact": {
      "sha256": "69c8eb644cb9dd022af2d574d7c1ecd62abbc51d1abb36db3a3e642b86ae4738",
      "format": "base",
      "quantization_namespace": "basert",
      "quantization_scheme": "base_q4"
    }
  },
  "runtime": {
    "name": "basert",
    "version": "0.2.6"
  },
  "benchmark": {
    "mode": "text",
    "chip": "NVIDIA GB10",
    "backend": "cuda",
    "protocol": {
      "report_schema": "computearena-benchmark/1",
      "benchmark_schema": "basert-benchmark-harness/1",
      "telemetry_schema": "basert-telemetry/4",
      "throughput_schema": "computearena-throughput/3",
      "comparable": true,
      "context_isolation": "headline_then_per_prefill",
      "execution_layout": "headline_then_prefill_processes",
      "model_load_in_timing": false,
      "decode_initial_context_tokens": 1,
      "conditioning_schema": "basert-conditioning/1",
      "conditioning_mode": "warmup_only",
      "cooldown_enabled": false
    },
    "parameters": {
      "context_tokens": 16384,
      "decode_context_tokens": 1
    },
    "throughput": {
      "prefill": [
        {
          "tokens": 128,
          "tokens_per_second": 5926.140208354926
        },
        {
          "tokens": 256,
          "tokens_per_second": 9908.972186648432
        },
        {
          "tokens": 512,
          "tokens_per_second": 14987.698091827586
        },
        {
          "tokens": 1024,
          "tokens_per_second": 16486.298108705334
        },
        {
          "tokens": 2048,
          "tokens_per_second": 16136.950642088967
        },
        {
          "tokens": 4096,
          "tokens_per_second": 14590.267992704989
        },
        {
          "tokens": 8192,
          "tokens_per_second": 12442.742051290455
        },
        {
          "tokens": 16384,
          "tokens_per_second": 9850.924875743387
        }
      ],
      "decode": {
        "generated_tokens": 128,
        "tokens_per_second": 69.39550994401873
      }
    }
  }
}