gpt-oss 120B — B200 vs MI355X
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and MI355X (AMD CDNA 4) 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.
Throughput at 123 tok/s/user on gpt-oss 120B: B200 hits 20100 tok/s/GPU, MI355X hits 7338. Per-million costs land at $0.03 and $0.06 respectively. B200 is 106% cheaper per token; B200 delivers 174% more tok/s/GPU.
B200 / MI355X on gpt-oss 120B at 183 tok/s/user: 11303 / 2493 tok/s/GPU, $0.05 / $0.17 per million tokens. B200 is 246% cheaper per token; B200 delivers 353% more tok/s/GPU.
Toward the upper edge of the 63–304 tok/s/user interactivity band, at 244 tok/s/user on gpt-oss 120B: B200 runs 6923 tok/s/GPU at $0.08/M tokens, MI355X runs 1252 at $0.34/M. B200 is 329% cheaper per token; B200 delivers 453% more tok/s/GPU. (Numbers reflect the default 1k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B200:20099.5MI355X:7338.3 | B200:11302.8MI355X:2493.0 | B200:6923.0MI355X:1252.0 |
| Cost ($/M tok) | B200:$0.027MI355X:$0.056 | B200:$0.048MI355X:$0.165 | B200:$0.079MI355X:$0.337 |
| tok/s/MW | B200:9262448MI355X:2769176 | B200:5208646MI355X:940755 | B200:3190341MI355X:472436 |
| Concurrency | B200:~205MI355X:~31 | B200:~64MI355X:~9 | B200:~63MI355X:~27 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.