GLM 5/5.1 — B300 vs MI325X
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and MI325X (AMD CDNA 3) on GLM 5/5.1. 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.
Near the low end of the 17–34 tok/s/user interactivity band, at 21 tok/s/user on GLM 5/5.1: B300 runs 1311 tok/s/chip at $0.48/M tokens, MI325X runs 173 at $1.79/M. B300 is 274% cheaper per token; B300 delivers 659% more tok/s/chip.
Setting 26 tok/s/user as the target on GLM 5/5.1, B300 produces 1169 tok/s/chip ($0.54 per million tokens) and MI325X produces 101 ($3.04). B300 is 462% cheaper per token; B300 delivers 1054% more tok/s/chip.
At 30 tok/s/user interactivity on GLM 5/5.1, B300 delivers 1062 tok/s/chip at $0.59 per million tokens; MI325X delivers 62 tok/s/chip at $4.91. B300 is 729% cheaper per token; B300 delivers 1626% more tok/s/chip at this point. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:1311.3MI325X:172.9 | B300:1169.4MI325X:101.3 | B300:1062.1MI325X:61.6 |
| Cost ($/M tok) | B300:$0.479MI325X:$1.789 | B300:$0.540MI325X:$3.037 | B300:$0.592MI325X:$4.907 |
| tok/s/MW | B300:690176MI325X:102283 | B300:615484MI325X:59939 | B300:559015MI325X:36421 |
| Concurrency | B300:~256MI325X:~34 | B300:~194MI325X:~16 | B300:~149MI325X:~9 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.