Webart Design

There's a wave of cheap open AI models right now. One of them does most everyday work about as well as the big providers, for about a fifth of the price, and it's free to download. So should you switch and cut your bill? Check out what we think in the comments.

Cheaper AI models won't cut your costs on their own Cheaper open models like GLM 5.2 promise big savings. WebArt Design on why the real cost sits in the system around the model, and how to capture it safely.

Saw this land a few days ago and immediately thought of our clients.

Anthropic pointed 𝗖𝗹𝗮𝘂𝗱𝗲 𝗠𝘆𝘁𝗵𝗼𝘀 an AI security model at Symfony, the framework Laravel is built on, and it found 19 real vulnerabilities. Laravel runs a huge chunk of the business web, so there's a good chance it's sitting under something you use.

Here's the thing though. The bugs aren't the scary part. They'd been there for years, some of them, just never found. The scary part is how many apps out there have nobody watching for the fix once it ships.

For the stuff we build and maintain, this was a quiet week. We stay on top of these releases and patch them as part of the deal, so nothing to panic about. But if you've got something that was built once and left alone since, this is a good nudge to go ask who's actually keeping it up to date. Link's in the comments.

Claude Mythos Found Flaws Under Millions of Laravel Sites. A Wake-Up Call, Not a Disaster Anthropic's Claude Mythos found 19 vulnerabilities in Symfony, the framework behind millions of Laravel sites. Here is what decides whether yours is safe.