---
license: apache-2.0
language:
- en
base_model:
- sequelbox/Qwen3-4B-PlumEsper
pipeline_tag: text-generation
library_name: transformers
tags:
- text-generation-inference
---

# **Qwen3-4B-PlumEsper**

> PlumEsper is a merged language model created using MergeKit, combining the specialized strengths and general reasoning capabilities of Esper 3 4B and Shining Valiant 3 4B. Built on Qwen/Qwen3-4B as the base, the merge was performed using the DELLA merge method. This fusion brings together the expertise of two powerful models: ValiantLabs/Qwen3-4B-ShiningValiant3 and ValiantLabs/Qwen3-4B-Esper3, resulting in a balanced model with enhanced versatility across a wide range of reasoning tasks.

## Model Files

| File Name | Size | Precision |
|-----------|------|-----------|
| Qwen3-4B-PlumEsper.BF16.gguf | 8.05 GB | BF16 |
| Qwen3-4B-PlumEsper.F16.gguf | 8.05 GB | F16 |
| Qwen3-4B-PlumEsper.F32.gguf | 16.1 GB | F32 |
| Qwen3-4B-PlumEsper.Q2_K.gguf | 1.67 GB | Q2_K |
| Qwen3-4B-PlumEsper.Q3_K_L.gguf | 2.24 GB | Q3_K_L |
| Qwen3-4B-PlumEsper.Q3_K_M.gguf | 2.08 GB | Q3_K_M |
| Qwen3-4B-PlumEsper.Q3_K_S.gguf | 1.89 GB | Q3_K_S |
| Qwen3-4B-PlumEsper.Q4_K_M.gguf | 2.5 GB | Q4_K_M |
| Qwen3-4B-PlumEsper.Q4_K_S.gguf | 2.38 GB | Q4_K_S |
| Qwen3-4B-PlumEsper.Q5_K_M.gguf | 2.89 GB | Q5_K_M |
| Qwen3-4B-PlumEsper.Q5_K_S.gguf | 2.82 GB | Q5_K_S |
| Qwen3-4B-PlumEsper.Q6_K.gguf | 3.31 GB | Q6_K |
| Qwen3-4B-PlumEsper.Q8_0.gguf | 4.28 GB | Q8_0 |

## 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)