gpt-oss 120B — B200 vs GB200 NVL72 Performance per Dollar
Cost per million tokens of B200 (NVIDIA Blackwell) versus GB200 NVL72 (NVIDIA Blackwell) on gpt-oss 120B. Owning-hyperscaler TCO normalized by output tokens — performance per dollar across LLM workloads. Pick the more cost-efficient SKU at every target interactivity level. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
Push gpt-oss 120B to 152 tok/s/user and B200 lands at $0.04 per million tokens against GB200 NVL72's $0.03 — GB200 NVL72 pulls ahead by 52%.
B200: $0.11 per million tokens. GB200 NVL72: $0.08. Both at 245 tok/s/user on gpt-oss 120B, with GB200 NVL72 44% cheaper.
Toward the upper edge of the 59–431 tok/s/user interactivity band — at 338 tok/s/user — B200 runs $0.27 per million tokens on gpt-oss 120B while GB200 NVL72 runs $0.23. GB200 NVL72 is the cheaper choice by 19%. (Numbers reflect the default 1k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Chip pricing (owning hyperscaler): B200 $1.73/chip/hr · GB200 NVL72 $1.86/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | B200:$0.044GB200 NVL72:$0.029 | B200:$0.114GB200 NVL72:$0.079 | B200:$0.272GB200 NVL72:$0.229 |
| Concurrency | B200:~68GB200 NVL72:~3024 | B200:~37GB200 NVL72:~71 | B200:~11GB200 NVL72:~16 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.