Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and GB200 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
Setting 110 tok/s/user as the target on DeepSeekv4 Pro 0813 1.6T, B200 produces 34892 tok/s/chip ($0.01 per million tokens) and GB200 NVL72 produces 30489 ($0.02). B200 is 23% cheaper per token; B200 delivers 14% more tok/s/chip.
At 169 tok/s/user interactivity on DeepSeekv4 Pro 0813 1.6T, B200 delivers 7269 tok/s/chip at $0.07 per million tokens; GB200 NVL72 delivers 13108 tok/s/chip at $0.04. GB200 NVL72 is 68% cheaper per token; GB200 NVL72 delivers 80% more tok/s/chip at this point.
B200 posts 3932 tok/s/chip for $0.12 per million tokens at 227 tok/s/user on DeepSeekv4 Pro 0813 1.6T; GB200 NVL72 posts 4252 tok/s/chip for $0.12. Cost per token is essentially tied; GB200 NVL72 delivers 8% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:34891.6GB200 NVL72:30488.9 | B200:7268.7GB200 NVL72:13108.1 | B200:3932.2GB200 NVL72:4252.2 |
| Cost ($/M tok) | B200:$0.014GB200 NVL72:$0.017 | B200:$0.066GB200 NVL72:$0.039 | B200:$0.122GB200 NVL72:$0.122 |
| tok/s/MW | B200:20404460GB200 NVL72:16304242 | B200:4250711GB200 NVL72:7009654 | B200:2299544GB200 NVL72:2273893 |
| Concurrency | B200:~108GB200 NVL72:~338 | B200:~9GB200 NVL72:~43 | B200:~4GB200 NVL72:~5 |
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