gpt-oss 120B · GPU comparison

gpt-oss 120B — H200 vs MI300X

Head-to-head AI inference benchmark comparison of H200 (NVIDIA Hopper) and MI300X (AMD CDNA 3) on gpt-oss 120B. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.

At 92 tok/s/user interactivity on gpt-oss 120B, H200 delivers 3635 tok/s/GPU at $0.11 per million tokens; MI300X delivers 2051 tok/s/GPU at $0.15. H200 is 38% cheaper per token; H200 delivers 77% more tok/s/GPU at this point.

H200 posts 2018 tok/s/GPU for $0.20 per million tokens at 144 tok/s/user on gpt-oss 120B; MI300X posts 957 tok/s/GPU for $0.32. H200 is 64% cheaper per token; H200 delivers 111% more tok/s/GPU.

Throughput at 197 tok/s/user on gpt-oss 120B: H200 hits 1085 tok/s/GPU, MI300X hits 355. Per-million costs land at $0.36 and $0.87 respectively. H200 is 139% cheaper per token; H200 delivers 206% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)

View performance-per-dollar view →

Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
Metric
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Throughput (tok/s/gpu)
H200:3634.9MI300X:2050.9
H200:2018.4MI300X:956.8
H200:1085.1MI300X:355.0
Cost ($/M tok)
H200:$0.108MI300X:$0.148
H200:$0.195MI300X:$0.320
H200:$0.365MI300X:$0.872
tok/s/MW
H200:2101076MI300X:1145772
H200:1166719MI300X:534550
H200:627235MI300X:198311
Concurrency
H200:~64MI300X:~23
H200:~62MI300X:~7
H200:~24MI300X:~8

Inference Performance

Inference performance metrics across different models, hardware configurations, and serving parameters.