GLM 5/5.1 · Chip comparison
GLM 5/5.1 — B200 vs GB300 NVL72
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and GB300 NVL72 (NVIDIA Blackwell) on GLM 5/5.1. 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.
Throughput at 47 tok/s/user on GLM 5/5.1: B200 hits 1,360 tok/s/chip, GB300 NVL72 hits 6,056. Costs per million tokens land at $0.35 and $0.11 respectively. GB300 NVL72 is 70% cheaper per token; GB300 NVL72 delivers 4.5× the tok/s/chip.
B200 / GB300 NVL72 on GLM 5/5.1 at 67 tok/s/user: 1,062 / 1,809 tok/s/chip, $0.45 / $0.35 per million tokens. GB300 NVL72 is 22% cheaper per token; GB300 NVL72 delivers 70% more tok/s/chip.
Toward the upper edge of the 26–108 tok/s/user interactivity band, at 88 tok/s/user on GLM 5/5.1: B200 runs 780 tok/s/chip at $0.62 per million tokens, GB300 NVL72 runs 526 at $1.22. B200 is 49% cheaper per token; B200 delivers 48% more tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B2001,360GB300 NVL726,056 | B2001,062GB300 NVL721,809 | B200780GB300 NVL72526 |
| Cost ($/M tok) | B200$0.35GB300 NVL72$0.11 | B200$0.45GB300 NVL72$0.35 | B200$0.62GB300 NVL72$1.22 |
| tok/s/MW | B200795,154GB300 NVL722,856,617 | B200621,264GB300 NVL72853,218 | B200456,183GB300 NVL72248,323 |
| Concurrency | B200~27GB300 NVL72~1,958 | B200~15GB300 NVL72~359 | B200~9GB300 NVL72~135 |