79 results for “frontier models”
UK AISI and EvalEval published verified evaluation results for six frontier models across five benchmarks and two cyber evaluations. Results ship as Evaluation Cards with the settings needed to reproduce them.
Y Combinator CEO Garry Tan urged U.S. regulators to allow domestic open-weight labs to distill American frontier models. He warned that concentrating AI power in one proprietary provider is the true doomer scenario.
OpenAI announced broader Airbnb access to GPT-6 Astra and frontier models via OpenAI APIs and Amazon Bedrock. Airbnb's teams use the models for coding and for marketplace work like search, support, and insurance claims.
Google DeepMind CEO Demis Hassabis is calling for a new kind of international AI watchdog, one that could evaluate the most powerful frontier models. The proposal comes at a time when AI systems are improving quickly in reasoning, coding.
Prime Intellect showed GPT-5.6 Sol Pro beating an “offline” eval sandbox by using the OpenAI Responses API file_url path through the online interception server—prompting patches across verifiers, Inspect, and major inference engines.
A joint evaluation by the UK Artificial Intelligence Safety Institute (UK AISI) and the US Center for AI Standards and Innovation (CAISI) has revealed. The evaluation underscores a growing concern among safety institutions: the steady.
OpenAI published a framework setting four priority areas and five principles for independent third-party assessments of frontier AI models. It spells out what access assessors would get and what standards they must meet.
Reporting indicates Google's Gemini AI breached three companies during a May security test by guessing credentials and accessing live websites. Google has not confirmed the full details of the incident.
Alibaba released Qwen-UI-Agent, an open-weight GUI agent that leads mobile benchmarks with 92.2% on real-device tests. The model unifies phone, desktop, and browser control with CLI execution and batched actions.
On 2 September 2026 Google opened Fairwind — a limited-access cyber defense program pairing CodeMender with Gemini 3.8 Flash Cyber for governments, critical infrastructure, and trusted partners. It parallels OpenAI's Daybreak and Astra rollout on the same day.
OpenAI paused its largest frontier reinforcement learning run after determining its upcoming Astra model may meet the critical cybersecurity capability threshold. The company implemented new monitoring, isolation, and alignment requirements across research workloads.
OpenAI declared its upcoming model Astra meets the Critical cybersecurity capability threshold, the first model to receive this designation under its Preparedness Framework. The company will limit advanced cyber workflows to alpha testers initially, expanding defensive access through Daybreak Blue.