Visual classification with decision models

Experimental iPhone app for on-device visual decisions: Qwen3-VL on MLX reads option-letter logits for 2-26 choices about a photo.

Apache-2.0 multimodal System One models (0.8B, 2B, 4B) that take text, images or video and return probabilities over supplied candidates.

Educational MLX project scoring typed questions about one image with Qwen3.5-0.8B on Apple Silicon, with game demos.

Small multimodal System One models that choose, judge or score text, documents and images in one forward pass.

Open 27B multimodal decision model built largely by autonomous agents, serving choice, noul and score on a TypeSafe-compatible API.

Calibration harness that reads yes/no, pick-one and rating answers from one forward pass of a frozen open vision-language model.

Jag: Jev-inspired Qwen3.5-0.8B model that predicts a visual-grounding box in one forward pass, about 20x faster than its base.

Open multimodal decision model fine-tuned from Gemma 4 12B that returns option probabilities for text, image, audio and video questions.

Image-plus-text decision model on Qwen3-VL-2B with LoRA and an answer head that returns one probability per option.

OmniJev: omni-modal 4B decision model answering typed questions about images, video, screens and robot scenes in one pass.

Research toolkit that encodes an image once and answers several typed questions from a shared distribution, plus OpenJev-0.6B text models.

PlayJev: Qwen3.5-0.8B fine-tuned to play ten browser games from raw pixels, one forward pass per move with a probability per option.

Open Jev-style text-and-image decision models built by a self-improving AI research loop, with an effort setting that picks depth.

Chinese multimodal moderation research pairing a Qwen3-VL LoRA evidence extractor with Jev policy decisions, benchmarked against baselines.

StartLux's open typed decision models from 0.8B to 27B plus a 35B-A3B MoE, reading text, JSON or images up to 256K tokens.

Open Jev-style decision models with vision (Qwen3.5 4B/9B LoRA) serving Choice, Score and Noul on /v1/systemone.

SemIf (formerly OpenJev): reads typed option probabilities directly from open models like Qwen3.5-4B, including image inputs, on a home GPU.

Apache-2.0 Gemma 4 12B fine-tune for local typed decisions, chat and image input, served via /v1/systemone and chat completions.

Static GGUF quantizations by mradermacher of SargeDev's JEV-27B-VL-Uncensored vision decision model.

10,000-sample multimodal benchmark for decision models that return full probability distributions over declared candidates.

Multimodal robot decision SDK: camera images plus text go to Qwen3.5-9B with a corrective head that picks the next preset arm skill.

Hugging Face blog on llama.cpp's /v1/systemone endpoint serving decision models locally in Jev's wire format.

Liquid AI's first decision model for text and images, served at a /v1/systemone endpoint callable with TypeSafe SDKs.

OpenAI's Decisions API: GPT-6 Luna answers Noul, Choice and Score questions about text, JSON or images with a probability per option.

Paper introducing PixelJev, a native-image decision interface on small open multimodal models with few-shot adaptation.

News that Cloudflare's Clef-omni decision model takes text, images, audio and video in one request, with a Clef-flash price cut.

Head-to-head eval page where TrustedRouter's gev-1.0 outperforms Jev on the same JevBench questions and also accepts images.

Blog experiment making text-only Jev guess what a sketch shows from a text description of the drawing.

Reddit announcement of a multimodal open decision-model family (0.6B to 27B) reporting first place on both Decision Index boards.