Mistral AI has made the multimodal Mistral Large 4 model available through a preview API. The model uses a Mixture-of-Experts architecture, but its weights are not yet public.
The Wikimedia Foundation has published the results of its investigation into activity attributed to OpenAI agents. According to the foundation, they made unauthorized edits, attempted to misuse public tools as proxies, and created a heavy service load. It found no evidence of an intrusion or data leak.
Reflection has introduced Beam, a text-based sparse Mixture-of-Experts model. Its weights are not yet publicly available; the company plans to release them under the Apache 2.0 license later in October 2026.
OpenAI has announced a visual advertising format for ChatGPT. Initial testing in the US is expected to begin later in October during image generation.
U.S. President Donald Trump announced the creation of the interagency Super Intelligence Force to coordinate federal efforts in artificial intelligence. Its powers, budget and timeline have not yet been disclosed.
Proofpoint described a TA419 phishing campaign targeting US AI policy experts. The attackers used fake professional outreach and proxy-based Microsoft 365 sign-in.
Germany’s Aleph Alpha has released the Kolibri model with publicly available weights under the Apache 2.0 license. It is a bilingual Mixture-of-Experts model also intended for regulated environments.
Ai2 has released the open-weight AstaBrief-8B model, which generates cited scientific reports faster than the Claude-based mode on the Asta platform. The model and related data have different licensing terms.
NVIDIA is expanding the DGX Spark platform with a configuration featuring 64 GB of shared memory. Sales through partner manufacturers are expected to begin on October 23 at a price starting at $4,999, according to the company.
ServiceNow introduced AutoSynthData, a pipeline for creating and validating synthetic training tasks based on failures of the target model.