TypeSafe AI adds image input and 4x context to Jev model

TypeSafe AI says its current Jev model now accepts images, placed anywhere text could go and in any combination, according to the company's documentation.

In a separate post, the company says Jev now handles four times as much context as before, and that users are no longer required to limit context length, though it still recommends doing so. Details are in the TypeSafe AI model docs.

In a third post, TypeSafe AI says its next releases will focus on more intelligence, better stability, and features that support real deployment, pointing to a blog announcement for the jev-1-14 update. The post also says larger surprises will come later. The post does not describe specific stability or deployment features.