LLM EXPLORER 58,133 MODELS INDEXED

Cosmos Reason1 7B Bnb 4bit by unsloth

By unsloth · 33 downloads
HF Repo ↗ Copy URL

Cosmos Reason1 7B Bnb 4bit is an open-source language model by unsloth. Features: 7b LLM, VRAM: 6GB, Context: 125K, License: other, Quantized, LLM Explorer Score: 0.18.

  Arxiv:2503.15558   4-bit   4bit Base model:nvidia/cosmos-reaso... Base model:quantized:nvidia/co...   Bitsandbytes   Conversational   Cosmos Dataset:nvidia/cosmos-reason1-... Dataset:nvidia/cosmos-reason1-... Dataset:nvidia/cosmos-reason1-...   En   Endpoints compatible   Image-text-to-text   Nvidia   Quantized   Qwen2 5 vl   Region:us   Safetensors   Unsloth
Model Card on HF 🤗: https://huggingface.co/unsloth/Cosmos-Reason1-7B-bnb-4bit 

Cosmos Reason1 7B Bnb 4bit Parameters and Internals

LLM NameCosmos Reason1 7B Bnb 4bit
Repository 🤗https://huggingface.co/unsloth/Cosmos-Reason1-7B-bnb-4bit 
Base Model(s)  Cosmos Reason1 7B   nvidia/Cosmos-Reason1-7B
Model Size7b
Required VRAM6 GB
Updated2026-08-11
Maintainerunsloth
Model Typeqwen2_5_vl
Model Files  6.0 GB
Supported Languagesen
Quantization Type4bit
Model ArchitectureQwen2_5_VLForConditionalGeneration
Licenseother
Context Length128000
Model Max Length128000
Transformers Version4.52.3
Tokenizer ClassQwen2Tokenizer
Padding Token<|vision_pad|>
Vocabulary Size152064
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Cosmos Reason1 7B Bnb 4bit

Best Alternatives
Context / RAM
Downloads
Likes
OlmOCR 2 7B 1025 Mlx 8bit125K / 9.4 GB1722
OlmOCR 2 7B 1025 Mlx 4bit125K / 5.6 GB1261
OlmOCR 7B 0725 8bit125K / 9.4 GB975
Qwen2.5 VL 7B Instruct 4bit125K / 5.6 GB34864
Qwen2.5 VL 7B Instruct 8bit125K / 8.9 GB175917
Jedi 7B 1080p 4bit125K / 4.9 GB191
Jedi 7B 1080p 6bit125K / 6.9 GB191
UI TARS 1.5 7B 6bit125K / 6.9 GB2094
UI TARS 1.5 7B 4bit125K / 4.9 GB1612
Qwen2.5 VL 7B Instruct 6bit125K / 6.9 GB1073
Note: green Score (e.g. "73.2") means that the model is better than unsloth/Cosmos-Reason1-7B-bnb-4bit.