Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and GB300 NVL72 (NVIDIA Blackwell) on DeepSeekv4 Pro 0813 1.6T. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics β 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 β
B300 / GB300 NVL72 on DeepSeekv4 Pro 0813 1.6T at 90 tok/s/user: 21708 / 88840 tok/s/chip, $0.03 / $0.01 per million tokens. GB300 NVL72 is 300% cheaper per token; GB300 NVL72 delivers 309% more tok/s/chip.
Around the middle of the 25β284 tok/s/user interactivity band, at 155 tok/s/user on DeepSeekv4 Pro 0813 1.6T: B300 runs 6876 tok/s/chip at $0.09/M tokens, GB300 NVL72 runs 20218 at $0.03/M. GB300 NVL72 is 188% cheaper per token; GB300 NVL72 delivers 194% more tok/s/chip.
Setting 220 tok/s/user as the target on DeepSeekv4 Pro 0813 1.6T, B300 produces 3781 tok/s/chip ($0.17 per million tokens) and GB300 NVL72 produces 5755 ($0.11). GB300 NVL72 is 49% cheaper per token; GB300 NVL72 delivers 52% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:21707.6GB300 NVL72:88839.8 | B300:6876.3GB300 NVL72:20218.1 | B300:3780.6GB300 NVL72:5754.6 |
| Cost ($/M tok) | B300:$0.029GB300 NVL72:$0.007 | B300:$0.091GB300 NVL72:$0.032 | B300:$0.166GB300 NVL72:$0.112 |
| tok/s/MW | B300:11425049GB300 NVL72:41905567 | B300:3619125GB300 NVL72:9536841 | B300:1989779GB300 NVL72:2714439 |
| Concurrency | B300:~30GB300 NVL72:~515 | B300:~8GB300 NVL72:~83 | B300:~4GB300 NVL72:~11 |
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 β’ Drag to pan β’ Double-click to reset β’ Click a point to pin tooltip
DeepSeek V4 Pro 0813 1.6T 8K / 1K Total Tokens per $1 TCO vs. Interactivity