Testing Exaone4 32B Q5

Today as always, I was testing a new AI LLM Model that came recently in Huggingface, called Exaone4. Here are my results using a 32b q5.

And some info about the model:

The EXAONE 4.0 model series consists of two sizes: a mid-size 32B model optimized for high performance, and a small-size 1.2B model designed for on-device applications.

• Number of Parameters (without embeddings): 30.95B
• Number of Layers: 64
• Number of Attention Heads: GQA with 40-heads and 8-KV heads
• Vocab Size: 102,400
• Context Length: 131,072 tokens

According to the Exaone team, their claim in their paper: https://arxiv.org/pdf/2507.11407 is included this benchmark results:

I can test more queries by request but this is so far a preliminary result.

I did test it with a webpage creation of dashboards and charts and here a screenshot of it:

Not so bad here, it finished the task with 460 code lines in 4 minutes in a 24 vram nvidia gpu. But then here the LLM Model was from HuiHui not from the source (Exaone), so is it because HuiHui or ?

What do you guys think, did I prompt a difficult query? What is a better way to test it?

Did you try this LLm Model, let me know your thoughts in the comments.