OpenAI
73 articles in the last 30 days.
-
GPT-6.2 Sol vs. Claude Opus 5.5: Not as accurate, but still might be the better buy
In three hands-on agent, coding and data analysis tests, Sol is about half the list price of Opus 5.5 and far cheaper per task, though Opus 5.5 is slightly more accurate.
-
Combating adversarial distillation
OpenAI describes how it detects and disrupts unauthorized distillation of its models, and how it hardened protected reasoning. Attack success on extracting chain-of-thought fell from 54.0% to 0.2% in GPT-6.2 Astra.
-
Hebbia sees improvements in financial research with GPT-6 Astra
Hebbia reports that GPT-6 Astra beat the best alternative it tested by 13.6 points on citing required source types, with 99.5% of cited excerpts relevant. It also improved presentation output.
-
Can AI help scientists identify promising drug candidates?
OpenAI introduces MedChemBench, 99 medicinal chemistry problems from 117 papers and patents. GPT-6.2 Astra scored 84.5%, against 58.2% for GPT-5.6 Sol, with selecting compounds and predicting properties harder.
-
OpenAI spent $1.2 million finding nearly 1,000 Linux security flaws. Sam Altman says this is just the beginning.
OpenAI's automated pipeline reproduced 1,724 kernel vulnerabilities and generated working exploits for over half. Its reported numbers are not yet independently verified or fixed.