TypeSafe AI releases Jev 1.14 with images, box type and 262k token context
TypeSafe AI announced Jev 1.14, the latest version of its System One decision model, now live under the jev‑preview alias. The update adds three experimental “danger zone” features: image input, a new Box question type that returns a normalized bounding box, and an optional long‑context mode that raises token limits to 262 k. TypeSafe notes that none of these features are production‑ready and…
Key points
- Jev 1.14 adds image input, Box bounding‑box questions and an optional 262k token context window
- Images are limited to PNG/JPEG up to 16 MiB and billed as regular input tokens
- Price stays at $0.042 per million input tokens; output tokens remain free
The model keeps its pricing unchanged at $42 per billion input tokens ($0.042 per million), with output tokens free. By default Jev 1.14 supports 64 k tokens per request and 32 k per question; enabling the long‑context header lifts both limits. Images must be PNG or JPEG, up to 16 MiB each, and are billed as regular input tokens. The Box type always returns four edge values and a confidence score, even when the target is absent, so TypeSafe recommends pairing it with a Noul check. Internal tests showed the Box test caught 30 of 36 cropped targets.
The rollout plan moves jev‑preview to Jev 1.14 now, with jev‑latest pointing to it in a week, while older versions receive deprecation notices. TypeSafe’s claims about multilingual gains and small performance improvements are not backed by public benchmarks.
TypeSafe AI Releases Jev 1.14 With Images, Box Type and 262k Context
MarkTechPost · 15 October 2026
Loading the full article…
This text was published by MarkTechPost and written by Asif Razzaq. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- SB Intuitions releases Sarashina free beta; translation test shows natural Japanese but meaning shifts · 1 src
- Google launches Gemini 4 Argon model globally for Pro and Ultra subscribers with introductory API pricing · 28 src
- The Nordics and NVIDIA Show What’s Possible When a Region Commits to Advanced Computing · 1 src
- LFM2.5-24B-A2B: Built for fast on-device agents · 1 src
- Simon Willison prompts Claude Opus 5.5 Max to build a Three.js web game · 1 src
Comments
via GitHub Discussions