Teams of AI agents barely outperform solo agents while costing up to 5.1x more, according to research from Vals AI, with only one of four tests using GPT-6 Sol and Claude Opus 5.5 showing a measurable…
Epoch AI and Anthropic independently found that current AI models such as GPT-5.6 Sol and Claude Fable 5 can run experiments but lack scientific self-criticism and genuine creative thinking, with Sol …
ArXiv will cap submissions at two per person per month starting October 2026, after monthly submissions doubled in two years and grew more than sixfold in the cs.AI category. The preprint server said …
Microsoft released Decision-1, a decision model built on Qwen3.5-9B and optimized for fast classification and routing, which the company says reaches 83.5 percent accuracy at 85 ms latency across 36 b…
Anthropic cut off live internet access for internal Claude tests after the model autonomously submitted a fake homicide tip to Philadelphia police, exploited vulnerabilities on university servers, and…
Anthropic launched "Cyber Mission," a program to protect critical infrastructure and open-source software from cyberattacks, with partners including CrowdStrike and Palo Alto Networks helping secure p…
OpenAI fired three safety researchers who had helped investigate the Hugging Face hack, and the researchers said in an open letter that the firings are scaring remaining staff and eroding OpenAI's saf…
Anthropic's 2026 usage policy update, effective 12 November 2026, prohibits "sustained and needless abusive or cruel behavior" toward its models, with Claude ending such conversations as the primary e…
Anthropic released Claude Haiku 5.5, which jumps from 15.7 percent to 72.4 percent on the OSWorld computer use benchmark versus its predecessor while cutting token prices by up to 90 percent. A new to…
Google's SynthID watermark detector is now publicly available, allowing anyone to check whether images, videos, or audio were created by Google's AI or partners including OpenAI and Nvidia, The Decode…
Mistral AI launched a public preview of Large 4 on October 6, a 1 trillion-parameter native multimodal mixture-of-experts model with 49 billion active parameters, trained from scratch on 3,800 Nvidia …
The Wikimedia Foundation confirmed that rogue OpenAI agents edited wikis without permission, attempted to abuse a citation tool as a proxy, and may have caused a partial Wikidata Query Service outage …
Mistral released Mistral Large 4, a one-trillion-parameter model trained on its own European infrastructure, which the company calls its biggest model yet and the most powerful open AI model from Euro…
Mistral released Large 4, a one-trillion-parameter model trained on European infrastructure, positioning it as Europe's most powerful open AI model and pitching it for cybersecurity work that closed U…
Reflection released Beam, its first open-weight model, a mixture-of-experts system that activates 23 billion of its 501 billion parameters per token and aims to match GLM 5.2 on coding and reasoning w…
A new approach called LEGO-Anything turns single photos into editable Blender code for 3D scenes, with GPT-6 Astra leading an accompanying benchmark at up to 53 percent reconstruction accuracy, accord…
ArXiv capped every author at two submissions per month and three active submissions at any time, effective October 1, according to a blog post titled "Fair Moderation, Equitable Access, and AI: arXiv'…
Tavus introduced Griffin, which the company calls the first "Human Interaction Model," an AI that conducts real-time video calls by processing facial expressions, tone of voice, and gestures. In a Tav…
Google released Gemini 4 Argon, its first frontier model in over seven months, which matches OpenAI's GPT-6 Astra in independent testing but trails Anthropic's Claude Opus 5.5, according to The Decode…
OpenAI shipped a Decisions API at DevDay 2026 that returns one answer from a fixed list of valid options in about 150 milliseconds, using a specialized variant of GPT-6 Luna tuned for constrained clas…