Jev 1.14.0: Images, Box, and 262k Context for the Decision-Making Model

TypeSafe AI released Jev 1.14.0 — an update to the 'System One' decision-making model featuring experimental image input, a new box question type, and a context of up to 262k tokens, while maintaining the $0.042 per 1M input tokens rate.

Compiled by Sergey KostenchukPublished 2026-10-15Updated 2026-10-15

2026-10-15 Coding

🛠 Jev 1.14.0: Images, Box, and 262k Context

TypeSafe AI released Jev 1.14.0 (alias jev-preview) — an update to the 'System One' decision-making model with three experimental features behind 'danger zone' headers: image input, box question type (object bounding box, confidence 0–1), and a context of up to 262k tokens instead of 64k. The rate remains the same: $0.042 per 1M input tokens, free output.

🌍 Affordable multimodal input and long context reduce the cost of document and image analysis, while box localizes objects without a separate vision model. The vendor warns: accuracy on images is lower than on text, and the response format may change.

👤 With a TypeSafe API key, images and context are enabled via the TypeSafe-Danger-Zone-Images and TypeSafe-Danger-Zone-Long-Context headers. In a week, jev-latest will move to 1.14.0, and jev-1.13.0 will be closed after a 60-day notice.

Source 1: https://typesafe.ai/blog/jev-1-14 Source 2: https://docs.typesafe.ai/images

Sources