gpt-oss 120B — B200 vs MI325X
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and MI325X (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.
Setting 72 tok/s/user as the target on gpt-oss 120B, B200 produces 25320 tok/s/GPU ($0.02 per million tokens) and MI325X produces 1999 ($0.18). B200 is 728% cheaper per token; B200 delivers 1167% more tok/s/GPU.
At 86 tok/s/user interactivity on gpt-oss 120B, B200 delivers 22213 tok/s/GPU at $0.02 per million tokens; MI325X delivers 1208 tok/s/GPU at $0.29. B200 is 1111% cheaper per token; B200 delivers 1739% more tok/s/GPU at this point.
B200 posts 20939 tok/s/GPU for $0.03 per million tokens at 100 tok/s/user on gpt-oss 120B; MI325X posts 835 tok/s/GPU for $0.43. B200 is 1542% cheaper per token; B200 delivers 2409% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B200:25319.6MI325X:1998.6 | B200:22212.9MI325X:1207.8 | B200:20938.8MI325X:834.5 |
| Cost ($/M tok) | B200:$0.022MI325X:$0.179 | B200:$0.024MI325X:$0.295 | B200:$0.026MI325X:$0.427 |
| tok/s/MW | B200:11668029MI325X:916809 | B200:10236368MI325X:554036 | B200:9649229MI325X:382814 |
| Concurrency | B200:~256MI325X:~60 | B200:~256MI325X:~31 | B200:~250MI325X:~17 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.