Newsroom · Updated daily
Model releases, research, funding, policy and public-sector AI — curated daily by the Indicops newsroom, with a dedicated desk for India.
Filtered
65 stories
Google DeepMind has released Gemini Robotics 2.5, its most capable vision-language-action model, tested on over 200 tasks and rolling out via Agile Robots, Apptronik and Boston Dynamics. It reached 96 per cent success packing automotive parts in Mercedes-Benz trials. AlphaGenome 2 launched the same day.
Anthropic has released Claude Fable 5.2 for general use and Mythos 5.2 for vetted cyber and life-sciences organisations. The company says output is up to 30 per cent faster than the 5.1 generation, at $8 per million input tokens and $40 per million output tokens, with lower rates of safeguard blocks.
Meta has released Muse Ember, the smallest model in its Muse family, aimed at high-volume agent work such as tool calls, extraction and routing. It offers a 1M-token context and 512K output tokens, priced from $0.10 per million input tokens, and is available through Meta's API and Google and Oracle clouds.
Google has expanded its eighth-generation TPU line with TPU 8tx, for confidential inference and small-scale training, and TPU 8i Max, a dual-chip package for long-context inference. It also open-sourced TorchTPU and TPU Sync. Availability dates for the new chips were not given.
NPR reports that a group calling itself Cyber Isnaad Front, which Israeli incident responders attribute to Iranian military hackers, has sabotaged industrial refrigeration systems from a food plant in Israel to roughly 300 French supermarkets and, in early November, retailers across six more European countries. Internet-exposed Danfoss controllers were a common entry point.
Meta has rolled out specialised versions of its models, trained with Muse Wonder, to every user of its Muse agent and Meta AI assistant. The company claims a 2.4-fold reduction in forgetting critical information and a 67 per cent improvement in completing everyday tasks that hit obstacles, with correct use of context rising from 76 to 90 per cent on its internal evaluation.
Nine months into the US–Israeli war on Iran, Washington carried out what President Trump called 'minor' strikes on Iranian air defences near the Strait of Hormuz on 6 November while Tehran reviewed a Qatari-mediated bridge proposal. A naval blockade, sanctions and near-record tanker attacks continue; no ceasefire is in force.
Brent crude fell more than 16 per cent to $98.32 in the week to 6 November as traders bet on a US–Iran agreement, even as tanker attacks in Hormuz continued at near-record pace. For AI operators the energy shock matters most through gas-linked electricity prices in Europe and Asia and through the cash Gulf states have left to invest.
After the midterm result, Democratic lawmakers and aides say they will introduce a broad set of bills on AI systems and data-centre expansion once they take control of Congress in January. Representative Lori Trahan wants independent experts to evaluate leading models against safety standards.
Bloomberg reports OpenAI has raised $11 billion so far at a $1.25 trillion valuation, with the round potentially reaching $20 billion. The figures fall below the $30 billion at $1.4 trillion reportedly sought in September, as investors grow wary of AI valuations and data-centre politics.
Google DeepMind says AI models working under human direction found two weaknesses in legacy 1024-bit RSA, covering signature forgery and cheaper batch factoring of related moduli. Keys of 2048 bits or more remain secure against this specific threat, according to the reporting.
Tencent's Hunyuan team has moved Hy4 from preview to general release, adding visual understanding and the ability to operate computers through graphical interfaces and command lines. The 770-billion-parameter mixture-of-experts model is live across Tencent Cloud TokenHub, Yuanbao, CodeBuddy and WeCom, as Chinese models dominate OpenRouter's token rankings.
