TypeSafe AI Releases Jev 1.14 With Images, Box Type and 262k Context

TypeSafe AI has shipped Jev 1.14, an update to its System One model published under the jev-preview alias. The release introduces three notable additions: experimental image input support, a Box type for bounding-box annotations, and an opt-in context window extended to 262k tokens.

The image input capability marks a step toward multimodal processing, allowing the model to accept visual data alongside text. The Box type suggests a design focus on structured visual reasoning, potentially useful for object detection or layout-aware tasks. Both features are labeled experimental, indicating they are still being refined and may change before a stable release.

The 262k-token context window is a significant expansion that could benefit long-document analysis, extended conversations, or codebase-level reasoning. Because it is opt-in, users can choose whether to enable it, likely balancing cost and latency against the need for longer context.

At $0.042 per million input tokens, the pricing positions Jev 1.14 competitively among API-based models. The combination of multimodal input, structured visual types, and extended context in a single update suggests TypeSafe AI is pushing toward a more versatile model capable of handling diverse input formats.

This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.