GLM 5.3 744B — H200 vs TPU7x
Head-to-head AI inference benchmark comparison of H200 (NVIDIA Hopper) 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 122 tok/s/user interactivity on GLM 5.3 744B, H200 delivers 2769 tok/s/chip at $0.12 per million tokens; TPU7x delivers 6896 tok/s/chip at $0.04. TPU7x is 195% cheaper per token; TPU7x delivers 149% more tok/s/chip at this point.
H200 posts 2306 tok/s/chip for $0.15 per million tokens at 134 tok/s/user on GLM 5.3 744B; TPU7x posts 6543 tok/s/chip for $0.04. TPU7x is 236% cheaper per token; TPU7x delivers 184% more tok/s/chip.
Throughput at 147 tok/s/user on GLM 5.3 744B: H200 hits 1545 tok/s/chip, TPU7x hits 6115. Per-million costs land at $0.22 and $0.05 respectively. TPU7x is 369% cheaper per token; TPU7x delivers 296% more tok/s/chip. (Numbers reflect the default agentic-traces · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H200:2768.5TPU7x:6895.6 | H200:2305.6TPU7x:6542.6 | H200:1545.1TPU7x:6114.6 |
| Cost ($/M tok) | H200:$0.122TPU7x:$0.041 | H200:$0.147TPU7x:$0.044 | H200:$0.219TPU7x:$0.047 |
| tok/s/MW | H200:2020825TPU7x:5713023 | H200:1682909TPU7x:5420512 | H200:1127835TPU7x:5065987 |
| Concurrency | H200:~12TPU7x:~8 | H200:~9TPU7x:~7 | H200:~5TPU7x:~6 |