Qwen 3.5 397B-A17B — B300 vs GB200 NVL72
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and GB200 NVL72 (NVIDIA Blackwell) 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 interactivity on Qwen 3.5 397B-A17B, B300 delivers 63294 tok/s/chip at $0.01 per million tokens; GB200 NVL72 delivers 52727 tok/s/chip at $0.01. GB200 NVL72 is 1% cheaper per token; B300 delivers 20% more tok/s/chip at this point.
B300 posts 43916 tok/s/chip for $0.01 per million tokens at 287 tok/s/user on Qwen 3.5 397B-A17B; GB200 NVL72 posts 24309 tok/s/chip for $0.02. B300 is 49% cheaper per token; B300 delivers 81% more tok/s/chip.
Throughput at 405 tok/s/user on Qwen 3.5 397B-A17B: B300 hits 7077 tok/s/chip, GB200 NVL72 hits 9186. Per-million costs land at $0.09 and $0.06 respectively. GB200 NVL72 is 58% cheaper per token; GB200 NVL72 delivers 30% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:63294.5GB200 NVL72:52727.4 | B300:43915.8GB200 NVL72:24309.2 | B300:7077.3GB200 NVL72:9185.8 |
| Cost ($/M tok) | B300:$0.010GB200 NVL72:$0.010 | B300:$0.014GB200 NVL72:$0.021 | B300:$0.089GB200 NVL72:$0.056 |
| tok/s/MW | B300:33312870GB200 NVL72:28196482 | B300:23113577GB200 NVL72:12999597 | B300:3724889GB200 NVL72:4912176 |
| Concurrency | B300:~41GB200 NVL72:~23 | B300:~39GB200 NVL72:~15 | B300:~4GB200 NVL72:~6 |