Devstral

What's the play for smaller base model training companies like Mistral ?

Mistral's positioning as the European alternative doesn't seem to be sticking. Acquisition seems tricky given how inflection, character.ai and stability have got carved out. The big acquisition bucks are going to product companies (windsurf)

They could pivot up the stack, but then they'd be starting from scratch with a team that's ill-suited for product development.

The base model offerings from pretraining companies have been surprisingly myopic. Deepmind seems to be the only one going past the obvious "content gen/coding automation" verticals. There's a whole world out there. LLM product companies are fast acquiring pieces of the real money pie and smaller pretraining companies are getting left out.

______

edit: my comment rose to the top. It's early in the morning. Injecting a splash of optimism.

LLMs are hard, and giants like Meta are struggling to make steady progress. Mistrals models are cheap, competent, open-source-ish and don't come with AGI-is-imminent baggage. Good enough for me.

To my own question: They have a list of target industries at the top. https://mistral.ai/solutions#industry

Good luck to them.

I actually think the underdog, cheap, but still capable independent european alternative is a decent selling point. They have also branched out into specialised models, and custom training. as well as their ocr service.