{
  "schema": "computearena-public-benchmark/1",
  "submission_id": "b13d950f-d732-4b69-bb9d-b27effdaf63c",
  "run_id": "3143cfc00fe87ac141a1a3ee715e46df",
  "created_at_unix_ms": 1791554629781,
  "submitted_by": "basecompute",
  "model": {
    "identity_resolution": "unresolved",
    "identity_verification": "unresolved",
    "name": "gemma-3-1b-it-cuda-q4mix",
    "architecture": "gemma3",
    "quantization": "base_q4",
    "artifact": {
      "sha256": "bf12127dd688be09db683c050cbe14f637382aab41d30e52752cdfd6910b2f34",
      "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": 14595.483908246568
        },
        {
          "tokens": 256,
          "tokens_per_second": 23251.335724914356
        },
        {
          "tokens": 512,
          "tokens_per_second": 33121.73969875617
        },
        {
          "tokens": 1024,
          "tokens_per_second": 39092.515560277585
        },
        {
          "tokens": 2048,
          "tokens_per_second": 38513.40563422221
        },
        {
          "tokens": 4096,
          "tokens_per_second": 36898.50604462388
        },
        {
          "tokens": 8192,
          "tokens_per_second": 36057.42699665739
        },
        {
          "tokens": 16384,
          "tokens_per_second": 34098.36629625011
        }
      ],
      "decode": {
        "generated_tokens": 128,
        "tokens_per_second": 265.13500685596824
      }
    }
  }
}