Hunyuan 4B Instruct is an open-source language model by tencent. Features: 4b LLM, VRAM: 8.5GB, Context: 256K, Instruction-Based, LLM Explorer Score: 0.21.
Base model:finetune:tencent/hu... Base model:tencent/hunyuan-4b-... Endpoints compatible Instruct Region:us Safetensors Sharded Tensorflow
Model Card on HF 🤗: https://huggingface.co/tencent/Hunyuan-4B-Instruct
Hunyuan 4B Instruct Benchmarks
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
What is the LLM Explorer Rank (Score)
Hunyuan 4B Instruct Parameters and Internals
| LLM Name | Hunyuan 4B Instruct |
| Repository 🤗 | https://huggingface.co/tencent/Hunyuan-4B-Instruct |
| Base Model(s) | |
| Model Size | 4b |
| Required VRAM | 8.5 GB |
| Updated | 2026-07-14 |
| Maintainer | tencent |
| Model Type | hunyuan_v1_dense |
| Instruction-Based | Yes |
| Model Files | |
| Model Architecture | HunYuanDenseV1ForCausalLM |
| Context Length | 262144 |
| Model Max Length | 262144 |
| Transformers Version | 4.41.2 |
| Tokenizer Class | PreTrainedTokenizerFast |
| Padding Token | <｜hy_▁pad▁｜> |
| Vocabulary Size | 120818 |
| Torch Data Type | bfloat16 |
Rank the Hunyuan 4B Instruct Capabilities
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟
What open-source LLMs or SLMs are you in search of? 55556 in total.
Original data from HuggingFace, Arena and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a