VibeThinker 3B is an open-source language model by WeiboAI. Features: 3b LLM, VRAM: 6.2GB, Context: 128K, License: mit, Instruction-Based, Code Generating, LLM Explorer Score: 0.53.
Arxiv:2606.16140 Base model:finetune:qwen/qwen2... Base model:qwen/qwen2.5-coder-... Code Codegen Conversational En Endpoints compatible Eval-results Gpqa Instruct Instruction-following Math Qwen2 Reasoning Region:us Safetensors Sharded Tensorflow
Model Card on HF 🤗: https://huggingface.co/WeiboAI/VibeThinker-3B
VibeThinker 3B 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)
VibeThinker 3B Parameters and Internals
| LLM Name | VibeThinker 3B |
| Repository 🤗 | https://huggingface.co/WeiboAI/VibeThinker-3B |
| Base Model(s) | |
| Model Size | 3b |
| Required VRAM | 6.2 GB |
| Updated | 2026-06-29 |
| Maintainer | WeiboAI |
| Model Type | qwen2 |
| Instruction-Based | Yes |
| Model Files | |
| Supported Languages | en |
| Generates Code | Yes |
| Model Architecture | Qwen2ForCausalLM |
| License | mit |
| Context Length | 131072 |
| Model Max Length | 131072 |
| Transformers Version | 4.51.3 |
| Tokenizer Class | Qwen2Tokenizer |
| Padding Token | <|endoftext|> |
| Vocabulary Size | 151936 |
| Torch Data Type | bfloat16 |
| Errors | replace |
Quantized Models of the VibeThinker 3B
Model |
Likes |
Downloads |
VRAM |
|---|---|---|---|
| Mythos Nano | 34 | 2900 | 1 GB |
| VibeThinker 3B 8bit | 3 | 2691 | 3 GB |
Best Alternatives to VibeThinker 3B
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Tessa T1 3B | 117K / 6.2 GB | 8 | 5 |
| UIGEN T1.5 3B | 117K / 6.2 GB | 10 | 1 |
| Qwen2.5 Coder 3B Instruct | 32K / 6.2 GB | 451109 | 101 |
| VisCoder2 3B | 32K / 6.2 GB | 972 | 3 |
| Qwen 2.5 3b Smarteaz 0.01a | 32K / 6.3 GB | 0 | 1 |
| Q2.5 CodeR1 3B | 32K / 6.2 GB | 0 | 1 |
| Qwen 2.5 Coder 3B SQL Writer | 32K / 6.2 GB | 3 | 1 |
| Omega Qwen2.5 Coder 3B | 32K / 6.2 GB | 15 | 3 |
| Arch Agent 3B | 32K / 6.2 GB | 29 | 8 |
| ....5 Coder 3B Instruct Finetuned | 32K / 6.2 GB | 10 | 0 |
Rank the VibeThinker 3B 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? 54653 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a