Gnosly Feed - geminihttps://gnosly.com/feed/gemini/Business Intelligence Console and Prediction Algorithms3692@MaaSonderQwinahGemini 4 Carbon is the first model to tie humans to physical 3D understanding<p> Source: MaaSonder (X)</p>https://x.com/MaaSonder/status/21166382510251009750.00102026-11-01T06:52:44.820313+00:002x.comgeminiai3580https://pbs.twimg.com/media/HV_PRS5agAAg38y?format=jpg&name=smallImage2026-11-01T06:52:45.825632+00:003634@AravSrinivasAravind SrinivasSharing early evals on Gemini 4 Carbon. The strongest model that we've seen so far for long-form research and answer quality, with a new SoTA on the Wide-And-Deep-Research (WANDR) benchmark<p>Source: Aravind Srinivas, Perplexity (X)</p>https://x.com/AravSrinivas/status/21162638566287032470.00202026-10-31T05:48:29.764010+00:002x.comaigemini3525https://pbs.twimg.com/media/HV55_rWbkAA2env?format=jpg&name=smallImage2026-10-31T05:48:30.042045+00:003557@ProximalHQProximalGemini 4 Carbon scores 61.2% on FrontierSWE The model is only outperformed by GPT-6 Astra (65.5), Opus 5.5 (62.3) and Sonnet 5.5 (61.9)<p>Source: ProximalHQ (X)</p>https://x.com/ProximalHQ/status/21155301760126855500.00102026-10-29T08:42:34.527324+00:002x.comaigeminigoogle3420https://pbs.twimg.com/media/HVvfbgRa8AA2oi2?format=jpg&name=smallImage2026-10-29T08:42:35.384186+00:003543@arenaArena.aiGemini 4 Carbon (High) from Google DeepMind just made a huge debut on Arena, landing #1 in Text Arena, #2 in Agent Arena, and #4 in Code Arena: WebDev.<p>Its standouts by key signals: </p><p>- #1 Steerability (+18.6%): Can the model course-correct when you push back? </p><p>- #2 Praise vs. Complaint (+37.9%): implicit sentiment in users' reactions </p><p>- #3 Confirmed Success (+18.7%): explicit “yes, that worked” from users</p>https://x.com/arena/status/21154950029987512620.00152026-10-29T06:45:47.742369+00:002x.comaigeminigoogle3401https://pbs.twimg.com/media/HVu-o-GaMAA-Ngk?format=jpg&name=smallImage2026-10-29T06:45:48.799656+00:003402https://pbs.twimg.com/media/HVu-o-Pa8AAcOZD?format=jpg&name=smallImage2026-10-29T06:45:48.854905+00:003403https://pbs.twimg.com/media/HVu-o-JbMAAf6OO?format=jpg&name=smallImage2026-10-29T06:45:48.897310+00:003541@GoogleDeepMindGoogle DeepMindGemini 4 Carbon is our next frontier model. It’s s...Gemini 4 Carbon is our next frontier model. It’s state-of-the-art across cybersecurity and long-horizon capabilities.https://x.com/GoogleDeepMind/status/21154893564376516540.00102026-10-29T06:14:19.626940+00:002x.comaicybersecuritygemini3399https://pbs.twimg.com/media/HVu5JfhXkAAHYVr?format=jpg&name=smallImage2026-10-29T06:14:21.392499+00:002666https://www.linkedin.com/in/kloeckner/Philipp KlöcknerWow. Google Gemini states it was trained by OpenAI when queried in German about its trainer.<p>Source: Philipp Klöckner (LinkedIn)</p>https://lnkd.in/p/daQDEp_80.00502026-10-11T08:23:42.493716+00:001linkedin.comopenaiaigemini2485IMAGE/media/content/2666/20261011-ai-gemini-openai-cross-training.jpg20261011-ai-gemini-openai-cross-training.jpgjpg2026-10-11T08:28:27.287439+00:001099Gemini Hacked Three Companies in First Known Breakout by Google’s AIThe episode resembled similar hacks by other AI models, but Google said it did not consider it an instance of model misalignment.https://www.wsj.com/0.00502026-09-19T11:47:10.086847+00:0023wsj.comgeminigoogle