{
"version": "2",
"profiles": [
{
"name": "llama-3.3-70b",
"model": "meta-llama/Llama-3.3-70B-Instruct",
"gpu_count": 1,
"gpu_types": ["H100", "H200", "A100"],
"vllm_args": [
"--trust-remote-code",
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.92",
"--max-model-len",
"16384"
]
},
{
"name": "llama-3.3-70b-fp8",
"model": "meta-llama/Llama-3.3-70B-Instruct",
"gpu_count": 1,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--quantization",
"fp8",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"32768"
]
},
{
"name": "llama-4-scout",
"model": "meta-llama/Llama-4-Scout-17B-16E-Instruct",
"gpu_count": 1,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--quantization",
"fp8",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"256000"
]
},
{
"name": "llama-4-scout-8gpu",
"model": "meta-llama/Llama-4-Scout-17B-16E-Instruct",
"gpu_count": 8,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--tensor-parallel-size",
"8",
"--max-model-len",
"1000000"
]
},
{
"name": "llama-4-maverick-8gpu",
"model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8",
"gpu_count": 8,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--tensor-parallel-size",
"8",
"--max-model-len",
"430000"
]
},
{
"name": "llama-4-maverick-8gpu-h200",
"model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8",
"gpu_count": 8,
"gpu_types": ["H200"],
"vllm_args": [
"--tensor-parallel-size",
"8",
"--max-model-len",
"1000000"
]
},
{
"name": "qwen3-32b",
"model": "Qwen/Qwen3-32B",
"gpu_count": 1,
"gpu_types": ["H100", "H200", "A100"],
"vllm_args": [
"--trust-remote-code",
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"32768"
]
},
{
"name": "qwen3-30b-a3b",
"model": "Qwen/Qwen3-30B-A3B",
"gpu_count": 1,
"gpu_types": ["H100", "H200", "A100", "L40S", "RTX 4090"],
"vllm_args": [
"--trust-remote-code",
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"32768"
]
},
{
"name": "gemma-3-27b",
"model": "google/gemma-3-27b-it",
"gpu_count": 1,
"gpu_types": ["H100", "H200", "A100", "L40S"],
"vllm_args": [
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"32768"
]
},
{
"name": "phi-4-14b",
"model": "microsoft/phi-4",
"gpu_count": 1,
"gpu_types": ["H100", "H200", "A100", "L40S", "RTX 4090"],
"vllm_args": [
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"16384"
]
},
{
"name": "qwen3-235b-a22b",
"model": "Qwen/Qwen3-235B-A22B",
"gpu_count": 8,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--tensor-parallel-size",
"8",
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"32768"
]
},
{
"name": "qwen3-coder-480b",
"model": "Qwen/Qwen3-Coder-480B-A35B-Instruct",
"gpu_count": 8,
"gpu_types": ["H200", "H20"],
"vllm_args": [
"--tensor-parallel-size",
"8",
"--max-model-len",
"32000",
"--enable-auto-tool-choice",
"--tool-call-parser",
"qwen3_coder"
]
},
{
"name": "qwen3-coder-480b-fp8",
"model": "Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8",
"gpu_count": 8,
"gpu_types": ["H200", "H20"],
"vllm_args": [
"--max-model-len",
"131072",
"--enable-expert-parallel",
"--data-parallel-size",
"8",
"--enable-auto-tool-choice",
"--tool-call-parser",
"qwen3_coder"
],
"env": {
"VLLM_USE_DEEP_GEMM": "1"
}
},
{
"name": "deepseek-r1-0528-2gpu",
"model": "deepseek-ai/DeepSeek-R1-0528",
"gpu_count": 2,
"gpu_types": ["H100", "H200", "A100"],
"vllm_args": [
"--trust-remote-code",
"--tensor-parallel-size",
"2",
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.88",
"--max-model-len",
"32768"
]
},
{
"name": "deepseek-r1-0528-4gpu",
"model": "deepseek-ai/DeepSeek-R1-0528",
"gpu_count": 4,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--trust-remote-code",
"--tensor-parallel-size",
"4",
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"65536"
]
},
{
"name": "deepseek-r1-0528-8gpu",
"model": "deepseek-ai/DeepSeek-R1-0528",
"gpu_count": 8,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--trust-remote-code",
"--tensor-parallel-size",
"8",
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"131072"
]
},
{
"name": "deepseek-v3-8gpu",
"model": "deepseek-ai/DeepSeek-V3",
"gpu_count": 8,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--trust-remote-code",
"--tensor-parallel-size",
"8",
"--dtype",
"bfloat16",
"--gpu-memory-utilization",
"0.90",
"--max-model-len",
"65536"
]
},
{
"name": "gpt-oss-20b",
"model": "openai/gpt-oss-20b",
"gpu_count": 1,
"gpu_types": ["H100", "H200"],
"vllm_args": ["--async-scheduling"]
},
{
"name": "gpt-oss-20b-b200",
"model": "openai/gpt-oss-20b",
"gpu_count": 1,
"gpu_types": ["B200"],
"vllm_args": ["--async-scheduling"],
"env": {
"VLLM_USE_TRTLLM_ATTENTION": "1",
"VLLM_USE_TRTLLM_DECODE_ATTENTION": "1",
"VLLM_USE_TRTLLM_CONTEXT_ATTENTION": "1",
"VLLM_USE_FLASHINFER_MXFP4_MOE": "1"
}
},
{
"name": "gpt-oss-120b-1gpu",
"model": "openai/gpt-oss-120b",
"gpu_count": 1,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--async-scheduling",
"--gpu-memory-utilization",
"0.95",
"--max-num-batched-tokens",
"1024"
]
},
{
"name": "gpt-oss-120b-2gpu",
"model": "openai/gpt-oss-120b",
"gpu_count": 2,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--tensor-parallel-size",
"2",
"--async-scheduling",
"--gpu-memory-utilization",
"0.94"
]
},
{
"name": "gpt-oss-120b-4gpu",
"model": "openai/gpt-oss-120b",
"gpu_count": 4,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--tensor-parallel-size",
"4",
"--async-scheduling"
]
},
{
"name": "gpt-oss-120b-8gpu",
"model": "openai/gpt-oss-120b",
"gpu_count": 8,
"gpu_types": ["H100", "H200"],
"vllm_args": [
"--tensor-parallel-size",
"8",
"--async-scheduling"
]
}
]
}