GraySoft
Projects Models Compare Cloud benchmarks FAQ Download guIDE →
Model Intelligence Sheet

prithivMLmods/OneJev1.5-4B-GGUF overview

OneJev1.5 4B GGUF OneJev1.5 4B https://huggingface.co/OmniJev/OneJev1.5 4B is a multimodal System One decision model fully fine tuned from Qwen/Qwen3.5 4B on a…

transformersgguftext-generation-inferencellama-cpponejevsystem-onedecision-modelcalibrationmultimodalgui-agentvideoimage-text-to-textenbase_model:OmniJev/OneJev1.5-4Bbase_model:quantized:OmniJev/OneJev1.5-4Blicense:apache-2.0endpoints_compatibleregion:usconversational

Runs locally from ~644.3 MB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).

Downloads
357
Likes
2
Pipeline
image-text-to-text

Repository Files & Downloads

10 GGUF files detected
Direct downloads for local inference
FileTypeQuantizationSizeLink
OneJev1.5-4B.BF16.ggufGGUFGGUF9.03 GBDownload
OneJev1.5-4B.Q3_K_L.ggufGGUFGGUF2.51 GBDownload
OneJev1.5-4B.Q3_K_M.ggufGGUFGGUF2.36 GBDownload
OneJev1.5-4B.Q4_K_M.ggufGGUFGGUF2.86 GBDownload
OneJev1.5-4B.Q4_K_S.ggufGGUFGGUF2.72 GBDownload
OneJev1.5-4B.Q5_K_M.ggufGGUFGGUF3.27 GBDownload
OneJev1.5-4B.Q5_K_S.ggufGGUFGGUF3.19 GBDownload
OneJev1.5-4B.Q6_K.ggufGGUFGGUF3.71 GBDownload
OneJev1.5-4B.Q8_0.ggufGGUFGGUF4.80 GBDownload
OneJev1.5-4B.mmproj-bf16.ggufGGUFBF16644.3 MBDownload

Model Details

Model IDprithivMLmods/OneJev1.5-4B-GGUF
AuthorprithivMLmods
Pipelineimage-text-to-text
Licenseapache-2.0
Base modelOmniJev/OneJev1.5-4B
Last modified2026-10-13T17:40:09.000Z

Model README

---

base_model:

  • OmniJev/OneJev1.5-4B

library_name: transformers

tags:

  • text-generation-inference
  • llama-cpp
  • onejev
  • system-one
  • decision-model
  • calibration
  • multimodal
  • gui-agent
  • video

license: apache-2.0

language:

  • en

pipeline_tag: image-text-to-text

---

OneJev1.5-4B-GGUF

> OneJev1.5-4B is a multimodal System One decision model fully fine-tuned from Qwen/Qwen3.5-4B on approximately 320,000 questions drawn from real agent runs, multimedia data, and public decision datasets. Designed for efficient and precise decision-making, it accepts diverse inputs—including text, screenshots, photos, and videos—alongside typed questions, returning calibrated probabilities for every option in a single forward pass. The model is highly competitive among sub-8B parameter models, achieving a leading score of 54.89 on the public portion of the Decision Index 0.3 benchmark, where it excelled particularly in Language Understanding and Tools & Automation. Accessible via the qev Python library and built around TypeSafe's System One API, OneJev1.5-4B serves as a powerful foundation for developing advanced GUI agents and complex multimodal decision workflows.

Model Files

| File Name | Quant Type | File Size | File Link | Description |

|-----------|------------|-----------|-----------|-------------|

| OneJev1.5-4B.BF16.gguf | BF16 | 9.7 GB | Link | Full BF16 weights. Highest quality, largest file size. |

| OneJev1.5-4B.Q3_K_L.gguf | Q3_K_L | 2.69 GB | Link | Lower quality but usable, good for low RAM availability. |

| OneJev1.5-4B.Q3_K_M.gguf | Q3_K_M | 2.54 GB | Link | Low quality. |

| OneJev1.5-4B.Q4_K_M.gguf | Q4_K_M | 3.07 GB | Link | Good quality, default size for most use cases, recommended. |

| OneJev1.5-4B.Q4_K_S.gguf | Q4_K_S | 2.92 GB | Link | Slightly lower quality with more space savings, recommended. |

| OneJev1.5-4B.Q5_K_M.gguf | Q5_K_M | 3.51 GB | Link | High quality, recommended. |

| OneJev1.5-4B.Q5_K_S.gguf | Q5_K_S | 3.43 GB | Link | High quality, recommended. |

| OneJev1.5-4B.Q6_K.gguf | Q6_K | 3.99 GB | Link | Very high quality, near perfect, recommended. |

| OneJev1.5-4B.Q8_0.gguf | Q8_0 | 5.16 GB | Link | Extremely high quality, generally unneeded but max available quant. |

| OneJev1.5-4B.mmproj-bf16.gguf | mmproj-bf16 | 676 MB | Link | Multimodal projection file in BF16 format. Used for vision/language models. |

llama.cpp

LLM inference in C/C++ — https://github.com/ggml-org/llama.cpp

Run prithivMLmods/OneJev1.5-4B-GGUF with guIDE

Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.

Download guIDE → · Browse 524k+ models · Compare models

Source: Hugging Face · Compare models