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

Nitishsharma9/CyberNexus-14B-GGUF overview

CyberNexus 14B GGUF GGUF quantizations of Qwen3.6 14B A3B FableVibes, a 14B MoE model fine tuned on reasoning traces and strictly optimized for incredibly fast…

ggufqwen14bcybersecurityethical-hackingcode-completionfimlicense:apache-2.0endpoints_compatibleregion:usconversational

Runs locally from ~4.96 GB disk (8 GB VRAM class GPUs with llama.cpp / guIDE).

Downloads
0
Likes
0
Pipeline

Repository Files & Downloads

6 GGUF files detected
Direct downloads for local inference
FileTypeQuantizationSizeLink
FableVibes-14B-Q2_K.ggufGGUFQ2_K4.96 GBDownload
FableVibes-14B-Q3_K_M.ggufGGUFQ3_K_M6.30 GBDownload
FableVibes-14B-Q4_K_M.ggufGGUFQ4_K_M7.88 GBDownload
FableVibes-14B-Q5_K_M.ggufGGUFQ5_K_M9.18 GBDownload
FableVibes-14B-Q6_K.ggufGGUFQ6_K10.55 GBDownload
FableVibes-14B-Q8_0.ggufGGUFQ8_013.65 GBDownload

Model Details

Model IDNitishsharma9/CyberNexus-14B-GGUF
AuthorNitishsharma9
Pipeline
Licenseapache-2.0
Base model
Last modified2026-07-12T21:32:08.000Z

Model README

---

license: apache-2.0

tags:

  • qwen
  • 14b
  • gguf
  • cybersecurity
  • ethical-hacking
  • code-completion
  • fim

---

CyberNexus-14B-GGUF

GGUF quantizations of Qwen3.6-14B-A3B-FableVibes, a 14B MoE model fine-tuned on reasoning traces and strictly optimized for incredibly fast Python scripting, Fill-in-the-Middle (FIM) code completion, Ethical Hacking, and Cybersecurity operations.

Background

This model started as a highly capable base and was pruned down to ~14B active parameters, removing over half its expert capacity. A single QLoRA pass was then orchestrated entirely by an autonomous AI agent, utilizing ~4,600 raw reasoning traces from Claude Fable 5 to recover capabilities lost during pruning.

Rather than focusing strictly on agentic orchestration, this model serves as a general-purpose reasoning distill specifically tailored for offensive and defensive security contexts. The Fable CoT traces provide structured multi-step reasoning patterns from a frontier-class model, distilled into a footprint that can run on consumer hardware.

Core Capabilities:

  • Lightning Fast Python Scripting: Optimized to generate robust, production-ready Python tools in milliseconds.
  • 🛡️ Ethical Hacking & Cyber Security: Deep knowledge of vulnerability assessment, penetration testing patterns, and defensive engineering.
  • 🔄 Fill-in-the-Middle (FIM): Native support for seamless code completion right inside your IDE.

Hardware compatibility

| Quantization | Bits | File Size (Est.) | RAM Required |

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

| Q2_K | 2-bit | ~5.32 GB | ~7 GB |

| Q3_K_M | 3-bit | ~6.77 GB | ~9 GB |

| Q4_K_M | 4-bit | ~8.47 GB | ~10.5 GB |

| Q5_K_M | 5-bit | ~9.85 GB | ~12 GB |

| Q6_K | 6-bit | ~11.3 GB | ~13.5 GB |

| Q8_0 | 8-bit | ~14.7 GB | ~17 GB |

Run Nitishsharma9/CyberNexus-14B-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