Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and B300 (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 37 tok/s/user on GLM 5/5.1: B200 hits 1631 tok/s/chip, B300 hits 1269. Per-million costs land at $0.29 and $0.49 respectively. B200 is 68% cheaper per token; B200 delivers 29% more tok/s/chip.
B200 / B300 on GLM 5/5.1 at 63 tok/s/user: 1121 / 941 tok/s/chip, $0.43 / $0.67 per million tokens. B200 is 56% cheaper per token; B200 delivers 19% more tok/s/chip.
Toward the upper edge of the 12–113 tok/s/user interactivity band, at 88 tok/s/user on GLM 5/5.1: B200 runs 780 tok/s/chip at $0.62/M tokens, B300 runs 683 at $0.92/M. B200 is 49% cheaper per token; B200 delivers 14% more tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:1631.4B300:1269.2 | B200:1121.2B300:941.3 | B200:780.1B300:683.0 |
| Cost ($/M tok) | B200:$0.295B300:$0.495 | B200:$0.429B300:$0.667 | B200:$0.616B300:$0.919 |
| tok/s/MW | B200:954015B300:667975 | B200:655649B300:495404 | B200:456183B300:359490 |
| Concurrency | B200:~216B300:~32 | B200:~17B300:~14 | B200:~9B300:~8 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr: VR200: 3.61H100: 1.17H200: 1.22B200: 1.73B300: 2.26GB200: 1.86GB300: 2.31MI300X: 0.95MI325X: 1.1MI355X: 1.5RTX6000PRO: 0.68Jalapeño (Teacup): 1.47
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