GLM 5.3 744B · Chip comparison
GLM 5.3 744B — B200 vs Vera Rubin NVL72
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and Vera Rubin NVL72 (NVIDIA Vera Rubin) 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 →
B200 / Vera Rubin NVL72 on GLM 5.3 744B at 236 tok/s/user: 7,942 / 138,484 tok/s/chip, $0.061 / $0.007 per million tokens. Vera Rubin NVL72 is 88% cheaper per token; Vera Rubin NVL72 delivers 17× the tok/s/chip.
Around the middle of the 204–333 tok/s/user interactivity band, at 268 tok/s/user on GLM 5.3 744B: B200 runs 7,081 tok/s/chip at $0.068 per million tokens, Vera Rubin NVL72 runs 97,953 at $0.010. Vera Rubin NVL72 is 85% cheaper per token; Vera Rubin NVL72 delivers 14× the tok/s/chip.
Setting 301 tok/s/user as the target on GLM 5.3 744B, B200 produces 6,183 tok/s/chip ($0.078 per million tokens) and Vera Rubin NVL72 produces 71,649 ($0.014). Vera Rubin NVL72 is 82% cheaper per token; Vera Rubin NVL72 delivers 12× the tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B2007,942Vera Rubin NVL72138,484 | B2007,081Vera Rubin NVL7297,953 | B2006,183Vera Rubin NVL7271,649 |
| Cost ($/M tok) | B200$0.061Vera Rubin NVL72$0.007 | B200$0.068Vera Rubin NVL72$0.010 | B200$0.078Vera Rubin NVL72$0.014 |
| tok/s/MW | B2004,644,264Vera Rubin NVL7241,964,979 | B2004,141,034Vera Rubin NVL7229,682,878 | B2003,615,968Vera Rubin NVL7221,711,755 |
| Concurrency | B200~4Vera Rubin NVL72~737 | B200~2Vera Rubin NVL72~424 | B200~1Vera Rubin NVL72~281 |