OpenAI Jalapeño chip posts first benchmarks
OpenAI published first-party benchmark numbers for Jalapeño: 1.5–1.9× more AI work per watt, 1.7–3.6× lower end-to-end latency. Vendor-tested, not independent.
OpenAI published first-party benchmark numbers for Jalapeño: 1.5–1.9× more AI work per watt, 1.7–3.6× lower end-to-end latency. Vendor-tested, not independent.