LLM EXPLORER 61,652 MODELS INDEXED

Synthia 4 27B by migtissera

By migtissera · 76 downloads
HF Repo ↗ Copy URL

Synthia 4 27B is an open-source language model by migtissera. Features: 27b LLM, VRAM: 55.6GB, License: apache-2.0, LLM Explorer Score: 0.33.

  Agentic   Assistant Base model:finetune:qwen/qwen3...   Base model:qwen/qwen3.8-27b   Conversational   Endpoints compatible   Image-text-to-text   Long-context   Mtp   Multimodal   Personal-ai   Qwen3   Qwen3 5   Reasoning   Region:us   Safetensors   Sharded   Synthia   Tensorflow   Tool-use
Model Card on HF 🤗: https://huggingface.co/migtissera/Synthia-4-27B 

Synthia 4 27B Parameters and Internals

LLM NameSynthia 4 27B
Repository 🤗https://huggingface.co/migtissera/Synthia-4-27B 
Base Model(s)  Qwen/Qwen3.8-27B   Qwen/Qwen3.8-27B
Model Size27b
Required VRAM55.6 GB
Updated2026-09-19
Maintainermigtissera
Model Typeqwen3_5
Model Files  4.0 GB: 1-of-18   3.0 GB: 2-of-18   2.5 GB: 3-of-18   4.0 GB: 4-of-18   2.1 GB: 5-of-18   4.0 GB: 6-of-18   2.1 GB: 7-of-18   4.0 GB: 8-of-18   2.1 GB: 9-of-18   4.0 GB: 10-of-18   2.1 GB: 11-of-18   4.0 GB: 12-of-18   2.1 GB: 13-of-18   4.0 GB: 14-of-18   2.1 GB: 15-of-18   4.0 GB: 16-of-18   2.1 GB: 17-of-18   3.4 GB: 18-of-18
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version5.8.0.dev0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to Synthia 4 27B

Best Alternatives
Context / RAM
Downloads
Likes
...wopus3.8 27B Flash Mlx OQ4e 1M1024K / 16.9 GB2111
Agens Pilot1024K / 54.7 GB1451
Qwen3.6 27B0K / 55.4 GB67281382197
Qwen3.8 27B0K / 55.6 GB28753
Qwen3.8 27B FP80K / 0.4 GB123157416
Qwen3.8 27B MLX0K / 16 GB141118
Qwen3.8 27B0K / 55.6 GB272721
Qwen3.6 27B NVFP40K / 22 GB1887525419
Qwen3.6 27B FP80K / 0.4 GB7740472342
Qwen3.5 27B0K / 55.2 GB25362941028
Note: green Score (e.g. "73.2") means that the model is better than migtissera/Synthia-4-27B.