GLM 5.3 744B · Chip comparison

GLM 5.3 744B — MI325X vs TPU7x

Head-to-head AI inference benchmark comparison of MI325X (AMD CDNA 3) and TPU7x (Google Ironwood) 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 →

At 33 tok/s/user interactivity on GLM 5.3 744B, MI325X delivers 1253 tok/s/chip at $0.24 per million tokens; TPU7x delivers 15933 tok/s/chip at $0.02. TPU7x is 1258% cheaper per token; TPU7x delivers 1172% more tok/s/chip at this point.

MI325X posts 1142 tok/s/chip for $0.27 per million tokens at 46 tok/s/user on GLM 5.3 744B; TPU7x posts 13353 tok/s/chip for $0.02. TPU7x is 1148% cheaper per token; TPU7x delivers 1069% more tok/s/chip.

Throughput at 59 tok/s/user on GLM 5.3 744B: MI325X hits 1052 tok/s/chip, TPU7x hits 10792. Per-million costs land at $0.29 and $0.03 respectively. TPU7x is 995% cheaper per token; TPU7x delivers 926% more tok/s/chip. (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.)

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)
MI325X:1252.7TPU7x:15933.4
MI325X:1142.4TPU7x:13353.5
MI325X:1052.1TPU7x:10792.4
Cost ($/M tok)
MI325X:$0.244TPU7x:$0.018
MI325X:$0.267TPU7x:$0.021
MI325X:$0.290TPU7x:$0.027
tok/s/MW
MI325X:741217TPU7x:13200857
MI325X:675949TPU7x:11063340
MI325X:622557TPU7x:8941529
Concurrency
MI325X:~3TPU7x:~31
MI325X:~2TPU7x:~21
MI325X:~1TPU7x:~16