BTCBTC
💲78785.51
-
0.9%
ETHETH
💲2454.58
-
1.22%
SOLSOL
💲96.99
-
4.54%
XRPXRP
💲1.44
-
4%
ZECZEC
💲757.96
-
6.39%
USDCUSDC
💲1.00
-
0%

Digg
Digg|8月 25, 2026 16:14
OpenAI says its first custom inference chip beat @nvidia's GB200 and GB300 systems on efficiency and latency. Across three open-weight models, @OpenAI's Jalapeño delivered 1.5 to 1.9x more work per watt. The company plans deployment by year-end.(Digg)
+3
Mentioned
Share To

Timeline

8月 26, 00:32OpenAI's Self-Developed Inference Chip Jalapeño First Round Testing
8月 26, 00:31Performance data of OpenAI's self-developed chip Jalapeño revealed for the first time
8月 26, 00:30The brand-new substreams-skills package offers professional expertise
8月 25, 23:41OpenAI's first self-developed inference chip Jalapeño surpasses NVIDIA GB300
8月 25, 15:25HIP-4 Testnet Clarifies Outcome Market Deployment Mechanism
8月 25, 14:53OpenAI plans to deploy Jalapeno chips
8月 25, 14:35OpenAI's self-developed AI inference chip Jalapeño outperforms NVIDIA.
8月 25, 14:05OpenAI releases Jalapeño chip test results
8月 25, 14:04OPENAI's Jalapeno chip outperforms NVIDIA in testing
8月 25, 04:05SpaceXAI will deploy NVIDIA Vera CPU

HotFlash

APP

X

Telegram

Facebook

Reddit

CopyLink

Hot Reads