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 132 tok/s/user on gpt-oss 120B: B200 hits 14063 tok/s/chip, MI355X hits 7454. Per-million costs land at $0.03 and $0.06 respectively. B200 is 63% cheaper per token; B200 delivers 89% more tok/s/chip.
B200 / MI355X on gpt-oss 120B at 194 tok/s/user: 7088 / 3059 tok/s/chip, $0.07 / $0.14 per million tokens. B200 is 103% cheaper per token; B200 delivers 132% more tok/s/chip.
Toward the upper edge of the 71–317 tok/s/user interactivity band, at 256 tok/s/user on gpt-oss 120B: B200 runs 3756 tok/s/chip at $0.13/M tokens, MI355X runs 1547 at $0.26/M. B200 is 106% cheaper per token; B200 delivers 143% more tok/s/chip. (Numbers reflect the default 1k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:14063.3MI355X:7454.3 | B200:7088.3MI355X:3058.7 | B200:3755.8MI355X:1546.6 |
| Cost ($/M tok) | B200:$0.034MI355X:$0.056 | B200:$0.067MI355X:$0.136 | B200:$0.127MI355X:$0.262 |
| tok/s/MW | B200:8224180MI355X:3566649 | B200:4145181MI355X:1463472 | B200:2196375MI355X:739983 |
| Concurrency | B200:~107MI355X:~29 | B200:~64MI355X:~8 | B200:~30MI355X:~7 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.