Qwen 3.5 397B-A17B — GB200 NVL72 vs H200
Head-to-head AI inference benchmark comparison of GB200 NVL72 (NVIDIA Blackwell) and H200 (NVIDIA Hopper) on Qwen 3.5 397B-A17B. 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 168 tok/s/user on Qwen 3.5 397B-A17B, GB200 NVL72 delivers 52727 tok/s/chip at $0.01 per million tokens; H200 hasn't been benchmarked at this target.
GB200 NVL72 hits 24309 tok/s/chip for $0.02 per million tokens at 287 tok/s/user on Qwen 3.5 397B-A17B. No H200 data at this operating point.
GB200 NVL72: 9186 tok/s/chip, $0.06 per million tokens at 405 tok/s/user on Qwen 3.5 397B-A17B. H200 is unmeasured here. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB200 NVL72:52727.4H200:— | GB200 NVL72:24309.2H200:— | GB200 NVL72:9185.8H200:— |
| Cost ($/M tok) | GB200 NVL72:$0.010H200:— | GB200 NVL72:$0.021H200:— | GB200 NVL72:$0.056H200:— |
| tok/s/MW | GB200 NVL72:28196482H200:— | GB200 NVL72:12999597H200:— | GB200 NVL72:4912176H200:— |
| Concurrency | GB200 NVL72:~23H200:— | GB200 NVL72:~15H200:— | GB200 NVL72:~6H200:— |