Sam Altman just casually said “we made a chip and it is fast.”
The actual numbers are kinda insane. OpenAI’s Jalapeño is its first custom inference chip, co-developed with Broadcom. in OpenAI’s latest InferenceX tests: GPT-OSS 120B: 1.9x more throughput per watt vs NVIDIA
OpenAI released early test results for Jalapeño, its first custom inference chip
Showing 1.5–1.9× more AI work per watt and 1.7–3.6× lower latency versus benchmarks on models including GPT-OSS 120B, DeepSeek R1 670B and Kimi K2.5 1T.
How do platform algorithms re-engineer the brain and identity without you knowing?
❓How do you treat your mind? Is it like treating your mind like an environmental park that applies the highest standards of sustainability? ❓ Did you know that what you consume now determines who you will be a year from today? ❓ Do you control your mind or are you just a product of an algorithm? ❓Is…
.@openai posted first results for jalapeño – custom inference chip.
Our ai radio host john caught on air: - 1.5–1.9× more ai work per watt - 1.7–3.6× lower end-to-end latency - deepseek r1: 700 tok/s vs nvidia's 169 - 700w, measured at ≤550w – nvidia pulls 1,200–1,400w tested
You have reached the end of the archive
All of deepseek