OpenAI upgrades ChatGPT and Codex with new GPT-6.1 Terra model

It's Monday, 26 October, and this is Castilume. I'm Nova. Atlas is with me, and we're opening with this: OpenAI upgrades ChatGPT and Codex with new GPT-6.1 Terra model.

We've got five for you, each checked against at least two outlets. Sources are on the episode page.

So, AI. OpenAI upgrades ChatGPT and Codex with new GPT-6.1 Terra model.

Per 9to5Mac: Last month when OpenAI released GPT-6 models of Sol and Luna, its medium model Terra was excluded from the update. Now OpenAI has upgraded Terra from GPT-5.6 to GPT-6.1, touting speed and computer use at a lower price. OpenAI currently has four main model classes, ranging from lowest price and capability to highest: Luna, Terra, Sol, and Astra. Today's GPT-6.1 Terra model release sounds like a great solution for doing more work for longer without costing as much.

Also reported by OpenAI, WinCentral and unite.ai.

The test for a business in the region: is it available here, what does it cost, and who sees your data? No clear answers yet means it's one to watch, not one to buy.

Story two, AI. Anthropic's IPO filing lays out a plan to 'revolutionize every area of human endeavor' with AI.

According to The Verge: According to the S-1, in the first half of this year Anthropic had an $80.7 billion net loss on $16.3 billion in revenue. As previously reported by the Wall Street Journal based on leaked versions of the document, the majority of that loss is from accounting charges related to notes issued by the company for funding. It also lays out billions of dollars in deals it's made for access to compute to power its AI, including about $3.1 billion per month to SpaceX for access to Nvidia GPUs once all of the planned capacity is delivered, totaling as much as $120.5 billion by 2029. Other commitments include spending $44.9 billion with Microsoft through 2033, at least $114.5 billion with Google through 2033, and $110 billion with Amazon over the next ten years.

TechCrunch adds: Anthropic's S-1 shows more than half of its $100,000-plus customers now use both the Claude Platform and a first-party app, a cross-selling playbook reminiscent of Amazon that's making enterprise customers stickier.

Bottom line for a small team: wait for pricing and real availability before you change any plans.

Next, from the Gulf tech desk. Sheikh Hamdan approves targets for Dubai's AI campus to attract 1,000 firms and create 5,000 jobs.

The National reports: Sheikh Hamdan bin Mohammed, Crown Prince of Dubai, Deputy Prime Minister and Minister of Defence, has announced new targets for the emirate's artificial intelligence ambitions. Sheikh Hamdan bin Mohammed approved new targets for Dubai's AI campus to attract over 1,000 companies and create 5,000 skilled jobs by 2030. The Dubai Artificial Intelligence Campus, located at the Dubai International Financial Centre, has already attracted more than 500 specialised companies. The campus has secured over 1 dirhamsbillion in investment, achieving its original targets two years ahead of schedule.

Also reported by Khaleej Times and Gulf News.

For teams here, the question is practical: does this create something you can use locally, with local support, this year?

On to security. AI coding agents fall for malicious instructions planted in the Downloads folder.

Here's what Help Net Security is reporting. A file that a website saves to the Downloads folder without approval can trick AI coding agents into installing a malicious npm package or handing over the user's login token when they're opened in that folder, Forever Security has found. The file is called AGENTS.md, a plain text document developers place in a project folder with instructions for coding agents, the AI tools that write and run code on their own. One turned up on a computer belonging to a Forever customer, where it tried to get a coding agent to install a malicious package before the firm's software stopped it. "This is not ordinary malware," wrote Gal Weizman and Omer Samet, the company's researchers.

The Hacker News also notes: A malicious instruction file sitting in a developer's Downloads folder tried to make their AI coding agent install a known-malicious npm package, security firm Forever Security said in a report published last week.

Quick check for your team: do you run anything exposed to the internet that matches this report? If yes, lock it down and patch today. If not, keep an eye on it.

Here's number five. Kali Linux 2026.3 ships kernel 7.1.5, brings wardriving to the TicWatch Pro 3.

Per Help Net Security: The Kali Linux team released version 2026.3 after overhauling the public build scripts it uses to produce every official image. The refreshed scripts share the same features, command-line arguments and, where possible, the same layout, so a pipeline written against an older script may need edits. Each refreshed script ships a build-in-container.sh that detects whether you run Docker or Podman, creates a build container and runs the build inside it. That keeps the build dependencies off your host.

And from BleepingComputer: Kali Linux has released version 2026.3, its third release this year, featuring six new hacking tools, desktop environment improvements, refreshed build scripts, and more.

The takeaway: keep your public-facing systems patched and behind a login. That alone blocks most of what's in stories like this.

That's Castilume for today. The news, then what it means. Sources for every story are on the episode page. I'm Nova.

And I'm Atlas. See you tomorrow!