Chinese cybersecurity firm 360 has launched Tulongfeng, an AI tool designed to compete with Anthropic’s Mythos, which is currently banned for export to non-Americans. Tokyo-based startup Sakana AI has introduced Fugu, a frontier AI model named after the Japanese blowfish.

The US government has granted Anthropic permission to release its AI model, Mythos, to a select group of trusted partners. This move allows specific US companies to access and utilize the capabilities of the Mythos model.

Anthropic has accused Alibaba of illegally extracting capabilities from its Claude AI model. The allegations suggest a breach of intellectual property rights related to AI technology.

The National Security Agency has lost access to a powerful AI model developed by Anthropic due to a dispute with the company. This loss deprives the NSA of a significant tool for cybersecurity analysis.

An elevated error rate has been reported across multiple AI models. Users can subscribe to receive updates via email or text message regarding incident resolutions and updates.

Mistral OCR 4 features bounding boxes, block classification, and inline confidence scores with extracted text. It supports 170 languages, operates in a single container for self-hosted deployments, and serves as an ingestion component for enterprise search and retrieval pipelines.

Unlimited-OCR is a new model designed for one-shot long-horizon parsing, building on the capabilities of Deepseek-OCR. It supports inference using Hugging Face transformers on NVIDIA GPUs and requires specific versions of Python and various libraries.

Major AI labs are expressing skepticism towards scaling models with endless parameters and training data. The US government banned Claude Fable 5 shortly after its release due to national security concerns stemming from a single jailbreak risk.

Anthropic employees expressed confusion and concern over the Trump administration's directive to remove their latest AI models due to national security issues. Executives were informed they had less than 90 minutes to comply with the order.

Europe can deploy a sovereign frontier-class AI model by utilizing existing public compute resources, as indicated by a sourced model and report. The region currently possesses tens of exaflops of public AI compute through EuroHPC supercomputers and national AI Factories, while new gigawatt datacenters face an average wait of 7.6 years for grid power.