---
license: apache-2.0
base_model:
- Vikhrmodels/QVikhr-3-4B-Instruction
language:
- en
- ru
pipeline_tag: text-generation
library_name: transformers
tags:
- ru
- text-generation-inference
---

# **QVikhr-3-4B-it-F32-GGUF**

> QVikhr-3-4B-Instruction, Instructive model based on Qwen/Qwen3-4B, trained on the Russian-language dataset GrandMaster2. Designed for high-efficiency text processing in Russian and English, delivering precise responses and fast task execution.

## Model Files

| File Name | Size | Type | Description |
|-----------|------|------|-------------|
| QVikhr-3-4B-Instruction.Q2_K.gguf | 1.67 GB | Model | Q2_K quantized model (smallest) |
| QVikhr-3-4B-Instruction.Q3_K_S.gguf | 1.89 GB | Model | Q3_K_S quantized model |
| QVikhr-3-4B-Instruction.Q3_K_M.gguf | 2.08 GB | Model | Q3_K_M quantized model |
| QVikhr-3-4B-Instruction.Q3_K_L.gguf | 2.24 GB | Model | Q3_K_L quantized model |
| QVikhr-3-4B-Instruction.Q4_K_S.gguf | 2.38 GB | Model | Q4_K_S quantized model |
| QVikhr-3-4B-Instruction.Q4_K_M.gguf | 2.5 GB | Model | Q4_K_M quantized model |
| QVikhr-3-4B-Instruction.Q5_K_S.gguf | 2.82 GB | Model | Q5_K_S quantized model |
| QVikhr-3-4B-Instruction.Q5_K_M.gguf | 2.89 GB | Model | Q5_K_M quantized model |
| QVikhr-3-4B-Instruction.Q6_K.gguf | 3.31 GB | Model | Q6_K quantized model |
| QVikhr-3-4B-Instruction.Q8_0.gguf | 4.28 GB | Model | Q8_0 quantized model |
| QVikhr-3-4B-Instruction.BF16.gguf | 8.05 GB | Model | BF16 precision model |
| QVikhr-3-4B-Instruction.F16.gguf | 8.05 GB | Model | F16 precision model |
| QVikhr-3-4B-Instruction.F32.gguf | 16.1 GB | Model | F32 full precision model (largest) |
| .gitattributes | 2.58 kB | Config | Git LFS configuration |
| config.json | 31 Bytes | Config | Model configuration |
| README.md | 149 Bytes | Documentation | Repository documentation |

## Quants Usage 

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):

![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)