What are the hardware requirements for Qwimi 3.6 27B Coder MTP BF16?

Qwimi 3.6 27B Coder MTP BF16 requires approximately 55.4 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwimi 3.6 27B Coder MTP BF16 and how large is it?

Qwimi 3.6 27B Coder MTP BF16 is developed by trjxter, a model with 27b 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 Qwimi 3.6 27B Coder MTP BF16?

Qwimi 3.6 27B Coder MTP BF16 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Qwimi 3.6 27B Coder MTP BF16 by trjxter — VRAM 55.4GB

Name: Qwimi 3.6 27B Coder MTP BF16
Author: trjxter

Qwimi 3.6 27B Coder MTP BF16 is an open-source language model by trjxter. Features: 27b LLM, VRAM: 55.4GB, License: apache-2.0, Code Generating, LLM Explorer Score: 0.25.

Agentic Base model:adapter:qwen/qwen3.... Base model:qwen/qwen3.6-27b Code Codegen Coding Conversational Dataset:jackrong/claude-opus-4... Dataset:jackrong/kimi-k2.5-rea... Dataset:minpeter/xlam-function... Dataset:nebius/swe-agent-traje... Dataset:swe-factory/deepswe-ag... Dataset:trjxter/kimi-k2.6-reas... Dataset:trjxter/kimi-k2.6-tech... Dataset:trjxter/kimi-k2.7-codi... En Endpoints compatible Image-text-to-text Lora Qwen3 5 Region:us Safetensors Sft Sharded Tensorflow Tool-calling Unsloth

Model Card on HF 🤗: https://huggingface.co/trjxter/Qwimi-3.6-27B-Coder-MTP-BF16

Qwimi 3.6 27B Coder MTP BF16 Parameters and Internals

LLM Name	Qwimi 3.6 27B Coder MTP BF16
Repository 🤗	https://huggingface.co/trjxter/Qwimi-3.6-27B-Coder-MTP-BF16
Model Name	unsloth/Qwen3.6-27B
Base Model(s)	Qwen/Qwen3.6-27B Qwen/Qwen3.6-27B
Model Size	27b
Required VRAM	55.4 GB
Updated	2026-09-02
Maintainer	trjxter
Model Type	qwen3_5
Model Files	4.0 GB: 1-of-15 3.9 GB: 2-of-15 3.9 GB: 3-of-15 3.9 GB: 4-of-15 3.9 GB: 5-of-15 3.9 GB: 6-of-15 4.0 GB: 7-of-15 3.9 GB: 8-of-15 3.9 GB: 9-of-15 3.9 GB: 10-of-15 3.9 GB: 11-of-15 3.9 GB: 12-of-15 4.0 GB: 13-of-15 3.9 GB: 14-of-15 0.5 GB: 15-of-15
Supported Languages	en
Generates Code	Yes
Model Architecture	Qwen3_5ForConditionalGeneration
License	apache-2.0
Model Max Length	262144
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|vision_pad\|>
Torch Data Type	bfloat16
Errors	replace

Rank the Qwimi 3.6 27B Coder MTP BF16 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 Qwimi 3.6 27B Coder MTP BF16

Best Alternatives	Context / RAM	Downloads	Likes
Qwopus3.6 27B Coder NVFP4	0K / 1.8 GB	505871	3
Veriloop Coder E1	0K / 55.6 GB	807	25
Damork Coder 27B Multimodal V2	0K / 54.7 GB	88	0
Damork Coder 27B Multimodal V2	0K / 54.7 GB	9	0
...d Fusion NEO CODER Mlx OQ4e 1M	1024K / 16.9 GB	190	1
Qwopus3.6 27B Coder 4bit	0K / 16 GB	312	1
Qwopus3.6 27B Coder 6bit	0K / 22.7 GB	101	1

Note: green Score (e.g. "73.2") means that the model is better than trjxter/Qwimi-3.6-27B-Coder-MTP-BF16.