Kimi K2.5/K2.6/K2.7-Code 1T — B200 vs Jalapeño
Head-to-head AI inference benchmark comparison of B200 (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.
At 91 tok/s/user interactivity on Kimi K2.5/K2.6/K2.7-Code 1T, B200 delivers 1746 tok/s/chip at $0.28 per million tokens; Jalapeño delivers 11420 tok/s/chip at $0.03. Jalapeño is 791% cheaper per token; Jalapeño delivers 554% more tok/s/chip at this point.
B200 posts 1004 tok/s/chip for $0.48 per million tokens at 122 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T; Jalapeño posts 9693 tok/s/chip for $0.04. Jalapeño is 1216% cheaper per token; Jalapeño delivers 866% more tok/s/chip.
Throughput at 153 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T: B200 hits 504 tok/s/chip, Jalapeño hits 7630. Per-million costs land at $0.95 and $0.05 respectively. Jalapeño is 1961% cheaper per token; Jalapeño delivers 1413% more tok/s/chip. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:1745.8Jalapeño:11419.7 | B200:1003.7Jalapeño:9692.7 | B200:504.4Jalapeño:7630.3 |
| Cost ($/M tok) | B200:$0.275Jalapeño:$0.031 | B200:$0.479Jalapeño:$0.036 | B200:$0.953Jalapeño:$0.046 |
| tok/s/MW | B200:1020964Jalapeño:10150852 | B200:586938Jalapeño:8615737 | B200:294951Jalapeño:6782530 |
| Concurrency | B200:~62Jalapeño:~1062 | B200:~7Jalapeño:~573 | B200:~18Jalapeño:~304 |