LLM EXPLORER 64,561 MODELS INDEXED

ThinkingCap Qwen3.8 27B W4A16 Merlin by halt95

By halt95 · 133 downloads
HF Repo ↗ Copy URL

ThinkingCap Qwen3.8 27B W4A16 Merlin is an open-source language model by halt95. Features: 27b LLM, VRAM: 0.2GB, License: other.

  Ampere   Awq Base model:bottlecapai/thinkin... Base model:quantized:bottlecap...   Compressed-tensors   Conversational   En   Fp8-kv   Gptq   Image-text-to-text   Int4   Long-context   Mtp   Qwen3.8   Qwen3 5   Region:us   Rtx-3090   Safetensors   Sm86   Speculative-decoding   Thinkingcap   Vision   Vllm   W4a16
Model Card on HF 🤗: https://huggingface.co/halt95/ThinkingCap-Qwen3.8-27B-W4A16-Merlin 

ThinkingCap Qwen3.8 27B W4A16 Merlin Parameters and Internals

LLM NameThinkingCap Qwen3.8 27B W4A16 Merlin
Repository 🤗https://huggingface.co/halt95/ThinkingCap-Qwen3.8-27B-W4A16-Merlin 
Base Model(s)  bottlecapai/ThinkingCap-Qwen3.8-27B   bottlecapai/ThinkingCap-Qwen3.8-27B
Model Size27b
Required VRAM0.2 GB
Updated2026-10-17
Maintainerhalt95
Model Typeqwen3_5
Model Files  0.0 GB   0.2 GB   16.8 GB
Supported Languagesen
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseother
Model Max Length262144
Transformers Version5.10.1
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to ThinkingCap Qwen3.8 27B W4A16 Merlin

Best Alternatives
Context / RAM
Downloads
Likes
...wopus3.8 27B Flash Mlx OQ4e 1M1024K / 16.9 GB2111
Agens Pilot1024K / 54.7 GB953
Qwen3.8 27B NPU2256K /  GB2524
JEV 27B VL0K / 55.6 GB1525286966
Clef0K / 54.7 GB201602194
Qwen3.6 27B0K / 55.4 GB67281382197
Qwen3.8 27B0K / 55.6 GB28753
Qwen3.8 27B0K / 55.6 GB272721
Qwen3.8 27B MLX0K / 16 GB141118
Qwen3.6 27B NVFP40K / 22 GB1887525419
Note: green Score (e.g. "73.2") means that the model is better than halt95/ThinkingCap-Qwen3.8-27B-W4A16-Merlin.