GLM 5.3 744B · Chip comparison

GLM 5.3 744B — B200 vs MI355X

Head-to-head AI inference benchmark comparison of B200 (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

Setting 108 tok/s/user as the target on GLM 5.3 744B, B200 produces 9909 tok/s/chip ($0.05 per million tokens) and MI355X produces 5478 ($0.08). B200 is 57% cheaper per token; B200 delivers 81% more tok/s/chip.

At 125 tok/s/user interactivity on GLM 5.3 744B, B200 delivers 8688 tok/s/chip at $0.06 per million tokens; MI355X delivers 4222 tok/s/chip at $0.10. B200 is 78% cheaper per token; B200 delivers 106% more tok/s/chip at this point.

B200 posts 7507 tok/s/chip for $0.06 per million tokens at 141 tok/s/user on GLM 5.3 744B; MI355X posts 1878 tok/s/chip for $0.22. B200 is 247% cheaper per token; B200 delivers 300% more tok/s/chip. (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.)

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/chip)
B200:9909.1MI355X:5477.6
B200:8688.3MI355X:4222.3
B200:7506.6MI355X:1877.7
Cost ($/M tok)
B200:$0.048MI355X:$0.076
B200:$0.055MI355X:$0.099
B200:$0.064MI355X:$0.222
tok/s/MW
B200:5794790MI355X:2620838
B200:5080886MI355X:2020234
B200:4389828MI355X:898399
Concurrency
B200:~15MI355X:~4
B200:~14MI355X:~2
B200:~11MI355X:~2

Inference Performance

Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.