LLM EXPLORER 59,420 MODELS INDEXED

Sailor2 20B by sail

By sail · 66 downloads
HF Repo ↗ Copy URL

Sailor2 20B is an open-source language model by sail. Features: 14b LLM, VRAM: 38.2GB, Context: 4K, License: apache-2.0, LLM Explorer Score: 0.16.

  Arxiv:2502.12982 Base model:finetune:qwen/qwen2...   Base model:qwen/qwen2.5-14b   Conversational   En   Endpoints compatible   Id   Jv   Km   Lo   Ms   Multilingual   My   Qwen2   Region:us   Safetensors   Sailor   Sea   Sharded   Su   Tensorflow   Th   Tl   Vi   Zh
Model Card on HF 🤗: https://huggingface.co/sail/Sailor2-20B 

Sailor2 20B Parameters and Internals

LLM NameSailor2 20B
Repository 🤗https://huggingface.co/sail/Sailor2-20B 
Base Model(s)  Qwen/Qwen2.5-14B   Qwen/Qwen2.5-14B
Model Size14b
Required VRAM38.2 GB
Updated2026-07-31
Maintainersail
Model Typeqwen2
Model Files  4.2 GB: 1-of-10   4.1 GB: 2-of-10   4.1 GB: 3-of-10   4.1 GB: 4-of-10   4.1 GB: 5-of-10   4.1 GB: 6-of-10   4.1 GB: 7-of-10   4.1 GB: 8-of-10   3.7 GB: 9-of-10   1.6 GB: 10-of-10
Supported Languagesen zh id th vi ms lo my jv km su tl
Model ArchitectureQwen2ForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version4.46.3
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151936
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Sailor2 20B

Best Alternatives
Context / RAM
Downloads
Likes
Qwen2.5 14B Instruct 1M986K / 29.6 GB10388342
T3Q Qwen2.5 14B V1.0 E3986K / 29.7 GB85228
ZYH LLM Qwen2.5 14B V4986K / 29.7 GB278
Qwen2.5 14B 1M YOYO V3986K / 29.7 GB2544
Qwen2.5 14B YOYO V4986K / 29.7 GB205
Qwen2.5 14B YOYO Latest V2986K / 29.7 GB1290
ZYH LLM Qwen2.5 14B V3986K / 29.7 GB238
...14B Instruct 1M GRPO Reasoning986K / 29.7 GB114
Etr1o V1.2986K / 29.7 GB110
Impish QWEN 14B 1M986K / 29.7 GB5626
Note: green Score (e.g. "73.2") means that the model is better than sail/Sailor2-20B.