gpt-oss 120B — B200 vs Jalapeño
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and Jalapeño (OpenAI Jalapeño) on gpt-oss 120B. 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.
Throughput at 177 tok/s/user on gpt-oss 120B: B200 hits 19679 tok/s/chip, Jalapeño hits 54802. Per-million costs land at $0.02 and $0.01 respectively. Jalapeño is 279% cheaper per token; Jalapeño delivers 178% more tok/s/chip.
B200 / Jalapeño on gpt-oss 120B at 258 tok/s/user: 12323 / 47075 tok/s/chip, $0.04 / $0.01 per million tokens. Jalapeño is 420% cheaper per token; Jalapeño delivers 282% more tok/s/chip.
Toward the upper edge of the 98–418 tok/s/user interactivity band, at 338 tok/s/user on gpt-oss 120B: B200 runs 5717 tok/s/chip at $0.08/M tokens, Jalapeño runs 35189 at $0.01/M. Jalapeño is 738% cheaper per token; Jalapeño delivers 515% more tok/s/chip. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:19679.2Jalapeño:54802.5 | B200:12323.3Jalapeño:47075.4 | B200:5717.2Jalapeño:35189.0 |
| Cost ($/M tok) | B200:$0.024Jalapeño:$0.006 | B200:$0.039Jalapeño:$0.007 | B200:$0.084Jalapeño:$0.010 |
| tok/s/MW | B200:11508326Jalapeño:48713326 | B200:7206622Jalapeño:41844786 | B200:3343400Jalapeño:31279154 |
| Concurrency | B200:~28Jalapeño:~820 | B200:~5Jalapeño:~397 | B200:~8Jalapeño:~181 |