Head-to-head AI inference benchmark comparison of GB300 NVL72 (NVIDIA Blackwell) and MI355X (AMD CDNA 4) 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 →
Near the low end of the 86–215 tok/s/user interactivity band, at 118 tok/s/user on GLM 5.3 744B: GB300 NVL72 runs 24241 tok/s/chip at $0.03/M tokens, MI355X runs 7254 at $0.06/M. GB300 NVL72 is 117% cheaper per token; GB300 NVL72 delivers 234% more tok/s/chip.
Setting 151 tok/s/user as the target on GLM 5.3 744B, GB300 NVL72 produces 15358 tok/s/chip ($0.04 per million tokens) and MI355X produces 3758 ($0.11). GB300 NVL72 is 165% cheaper per token; GB300 NVL72 delivers 309% more tok/s/chip.
At 183 tok/s/user interactivity on GLM 5.3 744B, GB300 NVL72 delivers 12528 tok/s/chip at $0.05 per million tokens; MI355X delivers 2796 tok/s/chip at $0.15. GB300 NVL72 is 191% cheaper per token; GB300 NVL72 delivers 348% more tok/s/chip at this point. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB300 NVL72:24241.4MI355X:7254.4 | GB300 NVL72:15357.8MI355X:3758.3 | GB300 NVL72:12528.0MI355X:2795.5 |
| Cost ($/M tok) | GB300 NVL72:$0.026MI355X:$0.057 | GB300 NVL72:$0.042MI355X:$0.111 | GB300 NVL72:$0.051MI355X:$0.149 |
| tok/s/MW | GB300 NVL72:11434613MI355X:3470987 | GB300 NVL72:7244236MI355X:1798224 | GB300 NVL72:5909446MI355X:1337570 |
| Concurrency | GB300 NVL72:~123MI355X:~10 | GB300 NVL72:~56MI355X:~2 | GB300 NVL72:~44MI355X:~4 |
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.
