GLM 5/5.1 · GPU comparison

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/GPU at $0.50/M tokens, MI325X runs 173 at $2.08/M. B300 is 320% cheaper per token; B300 delivers 659% more tok/s/GPU.

Setting 26 tok/s/user as the target on GLM 5/5.1, B300 produces 1169 tok/s/GPU ($0.56 per million tokens) and MI325X produces 101 ($3.53). B300 is 532% cheaper per token; B300 delivers 1054% more tok/s/GPU.

At 30 tok/s/user interactivity on GLM 5/5.1, B300 delivers 1062 tok/s/GPU at $0.61 per million tokens; MI325X delivers 62 tok/s/GPU at $5.71. B300 is 831% cheaper per token; B300 delivers 1626% more tok/s/GPU 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.)

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)
B300:1311.3MI325X:172.9
B300:1169.4MI325X:101.3
B300:1062.1MI325X:61.6
Cost ($/M tok)
B300:$0.496MI325X:$2.081
B300:$0.559MI325X:$3.534
B300:$0.613MI325X:$5.710
tok/s/MW
B300:604301MI325X:79293
B300:538903MI325X:46466
B300:489460MI325X:28234
Concurrency
B300:~256MI325X:~34
B300:~194MI325X:~16
B300:~149MI325X:~9

Inference Performance

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