Head-to-head AI inference benchmark comparison of GB200 NVL72 (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 67 tok/s/user on GLM 5/5.1: GB200 NVL72 hits 9135 tok/s/chip, GB300 NVL72 hits 10192. Per-million costs land at $0.06 and $0.06 respectively. GB200 NVL72 is 11% cheaper per token; GB300 NVL72 delivers 12% more tok/s/chip.
GB200 NVL72 / GB300 NVL72 on GLM 5/5.1 at 104 tok/s/user: 3619 / 4394 tok/s/chip, $0.14 / $0.15 per million tokens. GB200 NVL72 is 2% cheaper per token; GB300 NVL72 delivers 21% more tok/s/chip.
Toward the upper edge of the 30–177 tok/s/user interactivity band, at 141 tok/s/user on GLM 5/5.1: GB200 NVL72 runs 630 tok/s/chip at $0.82/M tokens, GB300 NVL72 runs 837 at $0.77/M. GB300 NVL72 is 7% cheaper per token; GB300 NVL72 delivers 33% more tok/s/chip. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB200 NVL72:9135.4GB300 NVL72:10192.4 | GB200 NVL72:3618.9GB300 NVL72:4393.6 | GB200 NVL72:629.8GB300 NVL72:837.2 |
| Cost ($/M tok) | GB200 NVL72:$0.057GB300 NVL72:$0.063 | GB200 NVL72:$0.143GB300 NVL72:$0.146 | GB200 NVL72:$0.820GB300 NVL72:$0.766 |
| tok/s/MW | GB200 NVL72:4885238GB300 NVL72:4807740 | GB200 NVL72:1935265GB300 NVL72:2072464 | GB200 NVL72:336816GB300 NVL72:394893 |
| Concurrency | GB200 NVL72:~698GB300 NVL72:~749 | GB200 NVL72:~231GB300 NVL72:~276 | GB200 NVL72:~62GB300 NVL72:~23 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr:
Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate cost per million tokens per decode chip or per prefill chip, rather than per total chip count. This makes direct token cost comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate input throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct input throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate output throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct output throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate power per decode chip or per prefill chip, rather than per total chip count. This makes direct power comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate Joules per decode chip or per prefill chip, rather than per total chip count. This makes direct Joules per token comparison with aggregated configs not an apples-to-apples comparison.
Matching measurements exist, but their chip series are hidden.
Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip
GLM5/5.1 744B 8K / 1K Total Tokens per $1 TCO vs. Interactivity