{
  "schema": "computearena-public-benchmark/1",
  "submission_id": "63c4ae2f-662c-4790-9240-8edd8755ab4e",
  "run_id": "66495b1a56d943ecae55344f3fa6fc87",
  "created_at_unix_ms": 1789838686976,
  "submitted_by": "prabod",
  "model": {
    "identity_resolution": "unresolved",
    "identity_verification": "unresolved",
    "name": "Llama-3.2-3B-Instruct-cuda-q4mix",
    "architecture": "llama",
    "quantization": "base_q4",
    "artifact": {
      "sha256": "356df54aa2a2731cc048312e5082ed89549f3a106a9a7583bcbac2d455e2c933",
      "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": 5907.587042660529
        },
        {
          "tokens": 256,
          "tokens_per_second": 9727.438344159375
        },
        {
          "tokens": 512,
          "tokens_per_second": 14651.421176217546
        },
        {
          "tokens": 1024,
          "tokens_per_second": 16486.46288251196
        },
        {
          "tokens": 2048,
          "tokens_per_second": 16166.617744778005
        },
        {
          "tokens": 4096,
          "tokens_per_second": 14683.110670656635
        },
        {
          "tokens": 8192,
          "tokens_per_second": 12586.22042766234
        },
        {
          "tokens": 16384,
          "tokens_per_second": 9806.507269272513
        }
      ],
      "decode": {
        "generated_tokens": 128,
        "tokens_per_second": 110.48396274122126
      }
    }
  }
}