Head-to-head AI inference benchmark comparison of B300 (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 39–215 tok/s/user interactivity band, at 83 tok/s/user on GLM 5.3 744B: B300 runs 18415 tok/s/chip at $0.03/M tokens, MI355X runs 10354 at $0.04/M. B300 is 18% cheaper per token; B300 delivers 78% more tok/s/chip.
Setting 127 tok/s/user as the target on GLM 5.3 744B, B300 produces 12839 tok/s/chip ($0.05 per million tokens) and MI355X produces 4196 ($0.10). B300 is 103% cheaper per token; B300 delivers 206% more tok/s/chip.
At 171 tok/s/user interactivity on GLM 5.3 744B, B300 delivers 7803 tok/s/chip at $0.08 per million tokens; MI355X delivers 3437 tok/s/chip at $0.12. B300 is 51% cheaper per token; B300 delivers 127% more tok/s/chip at this point. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:18414.8MI355X:10353.6 | B300:12838.8MI355X:4196.4 | B300:7803.0MI355X:3437.3 |
| Cost ($/M tok) | B300:$0.034MI355X:$0.040 | B300:$0.049MI355X:$0.099 | B300:$0.080MI355X:$0.121 |
| tok/s/MW | B300:9692023MI355X:4953873 | B300:6757260MI355X:2007862 | B300:4106855MI355X:1644627 |
| Concurrency | B300:~24MI355X:~8 | B300:~16MI355X:~2 | B300:~11MI355X:~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.
