25 September 2026

OpenAI prepares GPT-6 Cyber and its first product for deploying it

OpenAI will unveil GPT-6 Cyber, a model specialized in cybersecurity, in the coming weeks, Fortune reports, citing sources. It will debut alongside OpenAI's first product of its kind, meant to help customers deploy the model safely and fix vulnerabilities automatically. Selected participants of the Daybreak Red program are already testing the model.

Fortune ↗

From recent days

A third of the brands at the Paris Motor Show come from China - twice as many as two years ago

At the Paris Motor Show, which opens October 12, 20 of the roughly 60 exhibiting brands come from China, compared with 10 at the previous edition in 2024, Carscoops reports, citing Reuters. Alongside BYD, Geely, Chery, Jaecoo and Leapmotor, already known in Europe, Li Auto, Aito and Aion will also be there, among others; Chinese cars already account for more than 10 percent of new cars sold in Europe. Some of them are hybrids, which are not covered by the EU's anti-subsidy tariffs of up to 35 percent on Chinese EVs. European brands will bring the Audi A2 e-tron, Skoda Peaq, Dacia Striker and Fiat Grizzly, among others, while BMW is skipping the show.

Carscoops ↗

Reflection AI unveils Beam, an open 501-billion-parameter model said to match China's GLM-5.2 with 3-4 times less compute

Nvidia-backed US startup Reflection AI presented Beam on Monday, October 6 - a mixture-of-experts model with 501 billion parameters, 23 billion of them active at a time. According to the company, it matches China's GLM-5.2 in reasoning with 3-4 times less compute and comes close to Qwen 3.8-Max on some coding tasks, although in its own tests it still loses to some Chinese models. The model's weights, along with tools for running, evaluating and fine-tuning it, are due to be released in October; in the future, the company wants to sell an "AI factory" - software in which businesses build their own models on their own data instead of paying for ChatGPT or Claude. Reflection rents its compute from SpaceX: from July 1, 2026, through the end of 2029, it pays $150 million a month for Nvidia GB300 cards at the Colossus 2 data center - about $6.3 billion in total, although after the first three months either side can terminate the contract with 90 days' notice. Nvidia has invested $800 million in the startup.

Axios ↗ Axios (umowa ze SpaceX) ↗ Gizmodo ↗

Mistral Large 4 ranks highest in the Artificial Analysis index among models from outside the US and China - but five Chinese open models rank higher

Mistral released a public preview of Mistral Large 4, a trillion-parameter model (52 billion active) trained from scratch on 3,800 Nvidia Grace Blackwell chips in the company's European data centers; open weights are due by the end of October. According to independent firm Artificial Analysis, the model scored 38 points on its intelligence index, the same as GPT-6 Luna, and ranks highest among models from outside the US and China, but at least five Chinese open models score more, and an index task costs $1.13 with it, compared with $0.25 with GLM-5.3-Flash, Tom's Hardware reports. Large 4's strongest result is 82 percent on the CyberGym-E2E-AA test, the highest of all models; according to Mistral, Claude Opus 5.5 and GPT-6 Astra score close to zero there because they refuse to perform the task.

Mistral ↗ Tom's Hardware ↗

All news →