Kitani

Model · Text classification

OpenJudgement-4B-Preview

Kitani

Experimental open-weights judgment model by Kitani OpenJudgement is unfinished. We're releasing this checkpoint for people to experiment with, inspect, and build on. It still needs work on judgment quality, calibration, and inference efficiency. It is not as good as Jev overall in our internal task comparisons. It does show a meaningful improvement over untouched Qwen on our recorded validation comparison: 75.4% versus 64.2% annotation agreement. That is a result on a particular evaluation set, not a claim that we beat the base model on every task. There are questions it handles well and questions it confidently gets wrong. Please judge the preview by your own examples rather than assuming…

Model · Image and text to text

clover-1-150b-preview

Kitani

libraryname: transformers pipelinetag: image-text-to-text basemodel: - Qwen/Qwen3.8-27B - clover - mixture-of-experts - multimodal - reasoning - open-weights - custom-code Clover 1 150B Preview is the first public checkpoint of Clover 1, an experimental MoE model we're developing at Kitani. It's based on Qwen3.8-27B, but it isn't just a finetune with a different name. We kept much of Qwen3.8's underlying architecture, including the tokenizer, multimodal components, hybrid language backbone, embeddings, and LM head, while replacing the language model's dense FFNs with our own sparse mixture-of-experts setup. Clover currently has 8 full-sized experts per language layer with top-1 routing. The…

Dataset · Text classification

OpenJudgment-4B-Preview

Kitani

The custom human-supervised typed-judgment corpus used for OpenJudgement-4B-Preview, developed by Kitani. It adapts documented 2025–2026 source releases into Noul, Choice, and Score decisions. No Mix-v3 rows or local synthetic generators were reused in this corpus. This is an experimental training dataset, not evidence of Jev parity. Coverage is incomplete, source labels can be disputed, and the model remains unfinished. The source-specific terms below continue to apply. The selected model checkpoint was trained for 700 optimizer steps, seeing 44,741 decisions and 49,046,168 input tokens. Those are actual training exposures; the full corpus and split sizes below describe the available data…