Kimi K2.5/K2.6/K2.7-Code 1T — GB300 NVL72 vs Jalapeño
Head-to-head AI inference benchmark comparison of GB300 NVL72 (NVIDIA Blackwell) and Jalapeño (OpenAI Jalapeño) on Kimi K2.5/K2.6/K2.7-Code 1T. 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.
GB300 NVL72 posts 2056 tok/s/chip for $0.31 per million tokens at 91 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T; Jalapeño posts 11420 tok/s/chip for $0.03. Jalapeño is 910% cheaper per token; Jalapeño delivers 455% more tok/s/chip.
Throughput at 121 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T: GB300 NVL72 hits 1203 tok/s/chip, Jalapeño hits 9754. Per-million costs land at $0.53 and $0.04 respectively. Jalapeño is 1374% cheaper per token; Jalapeño delivers 711% more tok/s/chip.
GB300 NVL72 / Jalapeño on Kimi K2.5/K2.6/K2.7-Code 1T at 152 tok/s/user: 563 / 7707 tok/s/chip, $1.14 / $0.05 per million tokens. Jalapeño is 2389% cheaper per token; Jalapeño delivers 1269% more tok/s/chip. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB300 NVL72:2056.0Jalapeño:11419.7 | GB300 NVL72:1203.5Jalapeño:9754.4 | GB300 NVL72:563.1Jalapeño:7707.4 |
| Cost ($/M tok) | GB300 NVL72:$0.312Jalapeño:$0.031 | GB300 NVL72:$0.533Jalapeño:$0.036 | GB300 NVL72:$1.139Jalapeño:$0.046 |
| tok/s/MW | GB300 NVL72:969819Jalapeño:10150852 | GB300 NVL72:567684Jalapeño:8670601 | GB300 NVL72:265633Jalapeño:6851062 |
| Concurrency | GB300 NVL72:~92Jalapeño:~1062 | GB300 NVL72:~40Jalapeño:~584 | GB300 NVL72:~23Jalapeño:~311 |