Qwen 3.5 397B-A17B · Chip comparison
Qwen 3.5 397B-A17B — B300 vs H100
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and H100 (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 →
B300 posts 57,416 tok/s/chip for $0.011 per million tokens at 71 tok/s/user on Qwen 3.5 397B-A17B; H100 posts 6,053 tok/s/chip for $0.054. B300 is 80% cheaper per token; B300 delivers 9.5× the tok/s/chip.
Throughput at 123 tok/s/user on Qwen 3.5 397B-A17B: B300 hits 45,652 tok/s/chip, H100 hits 3,578. Costs per million tokens land at $0.014 and $0.091 respectively. B300 is 85% cheaper per token; B300 delivers 13× the tok/s/chip.
B300 / H100 on Qwen 3.5 397B-A17B at 176 tok/s/user: 22,215 / 2,528 tok/s/chip, $0.028 / $0.13 per million tokens. B300 is 78% cheaper per token; B300 delivers 8.8× the tok/s/chip. (Numbers reflect the default agentic-traces · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B30057,416H1006,053 | B30045,652H1003,578 | B30022,215H1002,528 |
| Cost ($/M tok) | B300$0.011H100$0.054 | B300$0.014H100$0.091 | B300$0.028H100$0.13 |
| tok/s/MW | B30030,218,900H1004,418,126 | B30024,027,170H1002,611,475 | B30011,691,981H1001,844,968 |
| Concurrency | B300~56H100~12 | B300~41H100~6 | B300~19H100~4 |