https://veriwire.news/daily-ai-news/can-an-open-model-do-security-research-cantinas--233634/ Veriwire en 2026-10-05T01:47:28.000Z Can an Open Model Do Security Research? Cantina's apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks https://veriwire.news/daily-ai-news/gpt-6-astra-vs-gpt-61-sol-vs-gemini-4-argon-vs-c-233635/ Veriwire en 2026-10-04T20:59:56.000Z GPT-6 Astra vs GPT-6.1 Sol vs Gemini 4 Argon vs Claude Fable 5.1: Which Frontier Model Fits Which Job https://veriwire.news/daily-ai-news/google-researchers-find-a-way-to-keep-self-impro-233331/ Veriwire en 2026-10-04T12:40:37.000Z Google researchers find a way to keep self-improving AI agents from memorizing their tests https://veriwire.news/daily-ai-news/nasa-and-ibms-open-source-lunar-model-turns-17-y-233332/ Veriwire en 2026-10-04T10:10:12.000Z NASA and IBM's open source lunar model turns 17 years of orbiter data into a foundation for lunar science https://veriwire.news/daily-ai-news/googles-new-gemini-tiers-cut-free-users-to-its-w-233334/ Veriwire en 2026-10-04T07:28:54.000Z Google's new Gemini tiers cut free users to its weakest model and lock $5/month subscribers out of Pro https://veriwire.news/daily-ai-news/aleph-alpha-releases-kolibri-a-781b-open-weight--233402/ Veriwire en 2026-10-04T07:01:04.000Z Aleph Alpha Releases Kolibri: A 78.1B Open-Weight English-German MoE Model With Only 3.46B Active Parameters https://veriwire.news/daily-ai-news/deepseek-harness-v02-brings-official-desktop-app-233403/ Veriwire en 2026-10-04T04:49:42.000Z DeepSeek Harness v0.2 Brings Official Desktop Apps to Its Open-Source Agent Harness https://veriwire.news/daily-ai-news/can-ai-read-pain-in-your-cats-face-what-the-rese-233215/ Veriwire en 2026-10-04T00:00:00.000Z Can AI read pain in your cat's face? What the research shows — and where chatbots fail https://veriwire.news/daily-ai-news/ai-made-starcraft-bot-swaps-in-human-made-bot-in-233687/ Veriwire en 2026-10-03T21:26:58.000Z AI Made StarCraft Bot Swaps In Human Made Bot In Tournament