Mistral Medium 3.5

Mistral just unveiled Medium 3.5, a new all-purpose AI model powering its chat app and cloud coding helpers that can keep working while you walk away. On paper, the pitch is simple: make the assistant smarter, let it handle longer jobs, and move those jobs off your laptop and into the cloud. But in the comments, the real show began: less “wow, cool launch,” more “fight me with benchmarks.”

The loudest reaction? A full-on debate over size, cost, and bragging rights. One camp says Mistral deserves credit for making something that can compete with much larger rivals without needing a ridiculous amount of hardware. That sparked hopeful “rooting for the underdog” energy, with fans cheering the company for betting that smaller, more practical models might beat giant expensive ones in real life. The other camp was not in a generous mood. Critics immediately dragged in rival models, basically saying, “Nice try, but this other one is faster, cheaper, and already runs great on my Mac.” Brutal.

Then came the product gripes. One commenter complained that Mistral’s web experience is so locked down that even simple website previews can break, which turned into a very relatable “why can’t I just test stuff without jumping through hoops?” mini-rant. So yes, Mistral launched a flashy new AI worker—but the crowd verdict is deliciously split: promising and practical to supporters, not quite king of the hill to skeptics, and mildly annoying to anyone who just wanted the demo to work.