Mistral OCR 4.1

Mistral just rolled out OCR 4.1, a new service that reads text from scanned pages and images, with extras like marking where paragraphs appear on a page and scoring how confident it is in what it read. On paper, it sounds like a serious upgrade for businesses dealing with piles of documents. In the comments, though, the launch got dragged straight into a full-blown AI identity crisis.

The loudest reactions were not gentle. One commenter bluntly declared that “the chinese did it better,” while another went full doom-post, saying they’d lost hope in Europe mattering in the AI race at all. So yes: what could have been a dry product update immediately became a geopolitical shouting match with a side of startup anxiety. That’s the vibe.

Then came the money complaints. At €3.5 per 1,000 pages, one user called it “expensive as hell” and basically said: if it’s not way better than free or older tools, why bother? Another commenter, speaking from the trenches of weird old books and fancy typography, shrugged that there was “nothing special” here for super-detailed work and claimed pricier rivals still win on quality. But not everyone showed up with a pitchfork: one calmer voice said Mistral’s tool is significantly faster than similar services in internal tests.

So the community verdict? Less “wow, cool feature set,” more “prove it, price it right, and stop acting like Europe’s still in the lead.” It’s half product launch, half comment-section cage match — and honestly, the cage match is the real event.