{
  "runs": [
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 186.47
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 59.68
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 129.43
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 61.28
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 57.73
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 25.29
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 14118.91
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 88.1
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 89.11
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 22.0
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1650.9257882903235
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 74.14
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 28.0
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 75.4
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 28.6
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1795.5908237208996
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 120.83
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 45.17
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1325.705122782343
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1878.7081268950026
    },
    {
      "model": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "model_clean": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "AWQ",
      "params_b": 80.0,
      "name_params_b": 80.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 680.226281561525
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1360.3779166371198
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 12436.56
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 84.04
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1095.31715306375
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 222.29
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 54.74
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 135.44
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 47.24
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 58.57
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 21.86
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 52928.91
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 61.4
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 984.5314161719857
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 108.91
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 47.9
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1019.986034633913
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 85.92
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 42.81
    },
    {
      "model": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "model_clean": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "AWQ",
      "params_b": 80.0,
      "name_params_b": 80.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 72209.01
    },
    {
      "model": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "model_clean": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "AWQ",
      "params_b": 80.0,
      "name_params_b": 80.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 71.26
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 104.59
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 38.86
    },
    {
      "model": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "model_clean": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "AWQ",
      "params_b": 80.0,
      "name_params_b": 80.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 519.96
    },
    {
      "model": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "model_clean": "cpatonn/Qwen3-Next-80B-A3B-Instruct-AWQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "AWQ",
      "params_b": 80.0,
      "name_params_b": 80.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 44.04
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 0.0
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 0.0
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 4078.97
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 78.49
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 80.34
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 37.1
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "default",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 859.7532189051877
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 93.13
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 47.81
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "default",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 2736.811464734795
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 123.92
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 73.89
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 118.15
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 54.56
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 86.17
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 28.21
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 14305.88
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 88.19
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 232.41
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 39.99
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1486.353457790027
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 97.06
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 32.9
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1890.096088342886
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 125.43
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 51.71
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1299.397329054036
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1575.9035072220947
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1429.9541207853842
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 3686.81
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 75.05
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1221.9380837920544
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 37309.58
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 113.18
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 235.33
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 55.13
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 80.83
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 19.52
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 0.0
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 0.0
    },
    {
      "model": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-27b-it-FP8-dynamic",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "FP8",
      "params_b": 27.0,
      "name_params_b": 27.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 621.3330493592667
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 104.14
    },
    {
      "model": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "model_clean": "RedHatAI/Qwen3-14B-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 14.0,
      "name_params_b": 14.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 41.41
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1024.3570195753498
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 81.58
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 37.28
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 127.38
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 45.15
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 1170.97
    },
    {
      "model": "openai/gpt-oss-20b",
      "model_clean": "openai/gpt-oss-20b",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 20.0,
      "name_params_b": 20.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 73.13
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 1.0",
      "tps_mean": 74.71
    },
    {
      "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "model_clean": "meta-llama/Meta-Llama-3.1-8B-Instruct",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "BF16",
      "params_b": 8.0,
      "name_params_b": 8.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 1.0",
      "tps_mean": 30.85
    },
    {
      "model": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "model_clean": "RedHatAI/gemma-3-12b-it-FP8-dynamic",
      "env": "TP1",
      "variant": "rocm",
      "gpu_config": "single",
      "quant": "FP8",
      "params_b": 12.0,
      "name_params_b": 12.0,
      "backend": "vLLM",
      "error": false,
      "test": "Throughput",
      "tps_mean": 1132.6153536237514
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TTFT @ QPS 4.0",
      "tps_mean": 194.04
    },
    {
      "model": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "model_clean": "cpatonn/Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
      "env": "TP2",
      "variant": "rocm",
      "gpu_config": "dual",
      "quant": "GPTQ",
      "params_b": 30.0,
      "name_params_b": 30.0,
      "backend": "vLLM",
      "error": false,
      "test": "TPOT @ QPS 4.0",
      "tps_mean": 58.93
    }
  ]
}