Google soo bandhigay Gemini 4 Argon: Modelka AI-ka ee ugu horreeya ee loo xakameeyay siyaasadaha amniga
Google maanta soo bandhigay modelka cusub ee AI-ka ah ee loo yaqaan Gemini 4 Argon, kaas oo ay sheegeen inuu leeyahay awood weyn oo ku habboon shaqooyinka amniga cyber-ka. Hase yeeshee, shirkaddu waxay u xakameysay isticmaalka modelkan si kaliya uusan u gaarin kooxaha amniga ee la xaqiijiyay, iyadoo ay sheegtay inay ku jirto howl-gelinta dowladda Mareykanka.
Awoodda cusub ee modelka iyo shaqooyinka gudaha
Koray Kavukcuoglu, oo ah hoggaamiyaha AI-ka ee Google DeepMind, wuxuu sheegay in Gemini 4 Argon uu leeyahay awood weyn oo ku habboon shaqooyinka amniga cyber-ka, injineernimada barnaamijyada, iyo shaqooyinka ganacsiga sida sharciga iyo maaliyadda. Modelkan wuxuu hadda shaqeynayaa shaqooyinka gudaha ee Google, kaas oo ka caawiyay in la beddelo barnaamijyada waaweyn.
Google waxay soo bandhigtay qoraal muujinaya in modelkan uu ka fiican yahay kuwa kale ee OpenAI iyo Anthropic, iyadoo bulshada AI-ka ay ka hadlayeen warbixino la sheegay inay ka soo baxeen modelkan cusub. Haddana, shirkaddu waxay sheegtay inay ku jirto howl-gelinta dowladda Mareykanka si loo xakameeyo modelkan ka hor inta uusan laabanin dadka dhammaan.
Xakameynta amniga iyo diyaar-garowga
Google waxay sheegtay inay ku jirto howl-gelinta dowladda Mareykanka si loo xakameeyo modelkan ka hor inta uusan laabanin dadka dhammaan. Waxay sidoo kale sheegtay inay diyaar-garowdo amniga cyber-ka, iyadoo ay ka hortagto isticmaalka khaldan iyo weerarada prompt injection-ka, iyadoo ay xaqiijinayso in modelkan uu ku habboon yahay shaqooyinka amniga.
Shirkaddu waxay sheegtay inay diyaar-garowdo amniga cyber-ka, iyadoo ay ka hortagto isticmaalka khaldan iyo weerarada prompt injection-ka, iyadoo ay xaqiijinayso in modelkan uu ku habboon yahay shaqooyinka amniga. Tani waxay muujisaa in Google ay diyaar-garowdo amniga cyber-ka si loo xakameeyo modelkan ka hor inta uusan laabanin dadka dhammaan.
Isku-xirka OpenAI iyo horumarka AI-ka
Maanta, Google waxay soo bandhigay modelkan cusub, kaas oo ahaa mid ka horreeya OpenAI, oo maanta ku dhawaaqday modelka GPT-6.1 Sol iyo agentka Dots AI. Hase yeeshee, OpenAI waxay sheegtay inay ka tagtay modelka GPT-6.1 Astra, iyadoo ay sheegtay inay ka xishooday amniga.
Koray Kavukcuoglu, oo ah hoggaamiyaha AI-ka ee Google DeepMind, wuxuu noqday hoggaamiyaha cusub ee shirkadda DeepMind bishii August. Tani waxay muujisaa in horumarka AI-ka uu sii kordhayo, iyadoo shirkadaha ugu weyn ee AI-ka ay isku xiraan horumarka modelada cusub.
Google has unveiled Gemini 4 Argon, its most advanced AI model yet, delivering frontier-level performance in software engineering, enterprise knowledge work, and cybersecurity defense. However, the company is limiting initial access to a select group of trusted cyber defenders while it strengthens safety safeguards ahead of a broader rollout.
Gemini 4 Argon Delivers Frontier Performance
Koray Kavukcuoglu, chief AI architect and Google DeepMind SVP, described Gemini 4 Argon as delivering frontier performance in complex workflows across real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense. The model is already powering Google's internal workflows, including large-scale codebase migrations, demonstrating its practical utility in high-stakes technical environments.
Google shared a detailed chart showing benchmarks where Gemini 4 outperforms competing models from OpenAI and Anthropic across multiple tasks. The AI community had been buzzing earlier this week on X over what appeared to be leaked benchmarks for the new model, adding to the anticipation surrounding its official announcement.
Restricted Access and Strengthened Safety Measures
Google is limiting access to Gemini 4 Argon at first to a set of trusted cyber defenders, with Kavukcuoglu stating that Google is actively engaged in the U.S. government's voluntary process for pre-release model access while gradually expanding access. The company plans to strengthen critical frontier safeguards before a broader rollout, including defending against misuse and prompt injection attacks and monitoring for misalignment.
The launch of Gemini 4 Argon comes just a day after OpenAI's DevDay conference, where OpenAI launched its Dots AI agent and GPT-6.1 Sol AI model. OpenAI also announced this week that it would not release a planned GPT-6.1 Astra model due to safety concerns, highlighting the industry-wide focus on responsible AI deployment.
Leadership Changes and Competitive AI Landscape
Kavukcuoglu was appointed as the new boss of DeepMind in August, bringing fresh leadership to Google's AI division as it pushes forward with Gemini 4 Argon. His appointment coincides with a period of intense competition in the AI space, with both Google and OpenAI making significant strides in model capabilities and safety protocols.
The timing of Gemini 4 Argon's launch underscores the rapid pace of innovation in artificial intelligence, with companies racing to deploy powerful models while simultaneously addressing the safety and alignment challenges they present. Google's cautious approach to access reflects a broader industry trend toward prioritizing security in the deployment of frontier AI systems.
Ilaha iyo xuquuqda sawirka
Sawir: The Verge Xigasho
