DeepSeek R1 — GB300 NVL72 vs Jalapeño
Head-to-head AI inference benchmark comparison of GB300 NVL72 (NVIDIA Blackwell) and Jalapeño (OpenAI Jalapeño) on DeepSeek R1. 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.
Near the low end of the 66–312 tok/s/user interactivity band, at 127 tok/s/user on DeepSeek R1: GB300 NVL72 runs 9494 tok/s/chip at $0.07/M tokens, Jalapeño runs 11045 at $0.03/M. Jalapeño is 112% cheaper per token; Jalapeño delivers 16% more tok/s/chip.
Setting 189 tok/s/user as the target on DeepSeek R1, GB300 NVL72 produces 2102 tok/s/chip ($0.31 per million tokens) and Jalapeño produces 7157 ($0.05). Jalapeño is 519% cheaper per token; Jalapeño delivers 240% more tok/s/chip.
At 251 tok/s/user interactivity on DeepSeek R1, GB300 NVL72 delivers 843 tok/s/chip at $0.76 per million tokens; Jalapeño delivers 2234 tok/s/chip at $0.16. Jalapeño is 382% cheaper per token; Jalapeño delivers 165% more tok/s/chip at this point. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB300 NVL72:9494.0Jalapeño:11044.9 | GB300 NVL72:2102.5Jalapeño:7156.7 | GB300 NVL72:842.8Jalapeño:2234.5 |
| Cost ($/M tok) | GB300 NVL72:$0.068Jalapeño:$0.032 | GB300 NVL72:$0.305Jalapeño:$0.049 | GB300 NVL72:$0.761Jalapeño:$0.158 |
| tok/s/MW | GB300 NVL72:4478308Jalapeño:9817707 | GB300 NVL72:991741Jalapeño:6361474 | GB300 NVL72:397565Jalapeño:1986203 |
| Concurrency | GB300 NVL72:~433Jalapeño:~653 | GB300 NVL72:~50Jalapeño:~220 | GB300 NVL72:~17Jalapeño:~39 |