Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and MI325X (AMD CDNA 3) on GLM 5.3 744B. 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.
AgentX replays real coding-agent sessions rather than fixed-length prompts, so context grows turn over turn and most of each request is served from cache instead of being recomputed. That turns the comparison into a systems question: KV transfer between nodes, prefix-aware routing, and cache capacity all move the curve alongside raw chip throughput. Fixed-sequence workloads stay the clean baseline for kernel and silicon performance, so the two scenarios answer different questions about the same hardware. Learn more about AgentX →
MI325X hits 813 tok/s/chip for $0.38 per million tokens at 8 tok/s/user on GLM 5.3 744B. No B300 data at this operating point.
MI325X: 813 tok/s/chip, $0.38 per million tokens at 14 tok/s/user on GLM 5.3 744B. B300 is unmeasured here.
At 21 tok/s/user on GLM 5.3 744B, MI325X delivers 775 tok/s/chip at $0.39 per million tokens; B300 hasn't been benchmarked at this target. (Numbers reflect the default agentic-traces · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:—MI325X:813.0 | B300:—MI325X:813.0 | B300:—MI325X:775.3 |
| Cost ($/M tok) | B300:—MI325X:$0.376 | B300:—MI325X:$0.376 | B300:—MI325X:$0.394 |
| tok/s/MW | B300:—MI325X:481083 | B300:—MI325X:481083 | B300:—MI325X:458746 |
| Concurrency | B300:—MI325X:~3 | B300:—MI325X:~3 | B300:—MI325X:~2 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
No measurements to plot for this selection. Review the benchmark controls above or adjust quick filters.
