{
  "schema": "computearena-public-benchmark/1",
  "submission_id": "fd2d8897-3178-4a2e-ba96-a82243c58367",
  "run_id": "1cad01652dd9932ec0107bd0f722e6df",
  "created_at_unix_ms": 1791559444132,
  "submitted_by": "basecompute",
  "model": {
    "identity_resolution": "unresolved",
    "identity_verification": "unresolved",
    "name": "gemma-4-26B-A4B-it-cuda-q8",
    "architecture": "gemma4",
    "quantization": "base_q4",
    "artifact": {
      "sha256": "6f8c9ebeab8106c4b4489ae19c41cec9bffcd9ed1aaf8489833bd365a86200d5",
      "format": "base",
      "quantization_namespace": "basert",
      "quantization_scheme": "base_q4"
    }
  },
  "runtime": {
    "name": "basert",
    "version": "0.3.0"
  },
  "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": 3193.583828543171
        },
        {
          "tokens": 256,
          "tokens_per_second": 4768.750202800561
        },
        {
          "tokens": 512,
          "tokens_per_second": 6005.015707914586
        },
        {
          "tokens": 1024,
          "tokens_per_second": 6494.120075343629
        },
        {
          "tokens": 2048,
          "tokens_per_second": 6253.284460160911
        },
        {
          "tokens": 4096,
          "tokens_per_second": 5912.601751662342
        },
        {
          "tokens": 8192,
          "tokens_per_second": 5355.533468976161
        },
        {
          "tokens": 16384,
          "tokens_per_second": 4371.684934203268
        }
      ],
      "decode": {
        "generated_tokens": 128,
        "tokens_per_second": 40.53737403455881
      }
    }
  }
}