LLM EXPLORER 58,178 MODELS INDEXED

SmallThinker 21BA3B Instruct by PowerInfer

By PowerInfer · 171 downloads
HF Repo ↗ Copy URL

SmallThinker 21BA3B Instruct is an open-source language model by PowerInfer. Features: 21b LLM, VRAM: 43GB, Context: 16K, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.21.

  Arxiv:2507.20984   En   Endpoints compatible   Instruct   Moe   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/PowerInfer/SmallThinker-21BA3B-Instruct 

SmallThinker 21BA3B Instruct Parameters and Internals

LLM NameSmallThinker 21BA3B Instruct
Repository 🤗https://huggingface.co/PowerInfer/SmallThinker-21BA3B-Instruct 
Model Namesmallthinker_21b_instruct
Model Size21b
Required VRAM43 GB
Updated2025-09-23
MaintainerPowerInfer
Instruction-BasedYes
Model Files  5.0 GB: 1-of-9   5.0 GB: 2-of-9   5.0 GB: 3-of-9   5.0 GB: 4-of-9   5.0 GB: 5-of-9   5.0 GB: 6-of-9   5.0 GB: 7-of-9   5.0 GB: 8-of-9   3.0 GB: 9-of-9   0.1 GB
Supported Languagesen
Model ArchitectureSmallThinkerForCausalLM
Licenseapache-2.0
Context Length16384
Model Max Length16384
Transformers Version4.53.3
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151936
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to SmallThinker 21BA3B Instruct

Best Alternatives
Context / RAM
Downloads
Likes
SmallThinker 21BA3B Instruct16K / 43 GB186112
Note: green Score (e.g. "73.2") means that the model is better than PowerInfer/SmallThinker-21BA3B-Instruct.