What are the hardware requirements for Keural Nova V1.2 Experimental?

Keural Nova V1.2 Experimental requires approximately 70.9 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Keural Nova V1.2 Experimental and how large is it?

Keural Nova V1.2 Experimental is developed by mkd-hossain, a model with 35b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Keural Nova V1.2 Experimental?

Keural Nova V1.2 Experimental is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Keural Nova V1.2 Experimental by mkd-hossain — VRAM 70.9GB

Name: Keural Nova V1.2 Experimental
Author: mkd-hossain

Keural Nova V1.2 Experimental is an open-source language model by mkd-hossain. Features: 35b LLM, VRAM: 70.9GB, License: apache-2.0, LLM Explorer Score: 0.24.

Agent Base model:finetune:qwen/qwen3... Base model:qwen/qwen3.6-35b-a3... Bilingual Conversational En Endpoints compatible Function-calling Image-text-to-text Keural Keural-nova Ko Korean Mkd Moe Qwen3 Qwen3 5 moe Region:us Safetensors Sharded Tensorflow Tool-calling

Model Card on HF 🤗: https://huggingface.co/mkd-hossain/Keural-Nova-v1.2-experimental

Keural Nova V1.2 Experimental Parameters and Internals

LLM Name	Keural Nova V1.2 Experimental
Repository 🤗	https://huggingface.co/mkd-hossain/Keural-Nova-v1.2-experimental
Base Model(s)	Qwen/Qwen3.6-35B-A3B Qwen/Qwen3.6-35B-A3B
Model Size	35b
Required VRAM	70.9 GB
Updated	2026-10-16
Maintainer	mkd-hossain
Model Type	qwen3_5_moe
Model Files	3.7 GB: 1-of-21 3.4 GB: 2-of-21 3.4 GB: 3-of-21 3.4 GB: 4-of-21 3.4 GB: 5-of-21 3.4 GB: 6-of-21 3.4 GB: 7-of-21 3.4 GB: 8-of-21 3.4 GB: 9-of-21 3.4 GB: 10-of-21 3.4 GB: 11-of-21 3.4 GB: 12-of-21 3.4 GB: 13-of-21 3.4 GB: 14-of-21 3.4 GB: 15-of-21 3.4 GB: 16-of-21 3.4 GB: 17-of-21 3.4 GB: 18-of-21 3.4 GB: 19-of-21 3.4 GB: 20-of-21 2.6 GB: 21-of-21 1.7 GB
Supported Languages	ko en
Model Architecture	Qwen3_5MoeForConditionalGeneration
License	apache-2.0
Model Max Length	262144
Transformers Version	5.12.1
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Errors	replace

Rank the Keural Nova V1.2 Experimental capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Keural Nova V1.2 Experimental

Best Alternatives	Context / RAM	Downloads	Likes
....5 35B A3B Sceptic V3 1M W4A16	1024K / 26.1 GB	223	0
...AgentWorld 35B A3B Abliterated	256K / 69.2 GB	962	18
Ornith 1.5 35B A3B NPU2	256K / GB	72	0
Qwen3.6 35B A3B NPU2	256K / GB	86	0
Qwen3.6 35B A3B	0K / 72.2 GB	5540855	2648
Qwen3.5 35B A3B	0K / 72.3 GB	2158511	1482
Qwen3.6 35B A3B NVFP4	0K / 23.4 GB	8788684	482
Ornith 1.0 35B	0K / 70.4 GB	1369192	516
Qwen3.6 35B A3B FP8	0K / 0.8 GB	8984811	343
Qwen AgentWorld 35B A3B	0K / 69.2 GB	66155	665

Note: green Score (e.g. "73.2") means that the model is better than mkd-hossain/Keural-Nova-v1.2-experimental.