45 lines
4.3 KiB
Markdown
45 lines
4.3 KiB
Markdown
---
|
||
license: mit
|
||
base_model:
|
||
- WeiboAI/VibeThinker-3B
|
||
language:
|
||
- en
|
||
pipeline_tag: text-generation
|
||
library_name: transformers
|
||
tags:
|
||
- math
|
||
- code
|
||
- reasoning
|
||
- gpqa
|
||
- instruction-following
|
||
- text-generation-inference
|
||
- llama-cpp
|
||
---
|
||
|
||
# **VibeThinker-3B-GGUF**
|
||
|
||
> **[VibeThinker-3B](https://huggingface.co/WeiboAI/VibeThinker-3B)** is a 3-billion-parameter reasoning-focused language model developed by WeiboAI, built on top of Qwen2.5-Coder-3B and trained using the Spectrum-to-Signal Principle (SSP) post-training pipeline, which combines curriculum-based two-stage SFT, multi-domain reinforcement learning via MaxEnt-Guided Policy Optimization (MGPO), offline self-distillation, and instruct RL to systematically develop strong verifiable reasoning capabilities across mathematics, coding, and STEM tasks. Motivated by the Parametric Compression-Coverage Hypothesis — which posits that verifiable reasoning is a highly compressible, parameter-dense capability that compact models can carry near-frontier performance in — VibeThinker-3B achieves remarkable results for its size, scoring 76.4 on IMO-AnswerBench (improving to 80.6 with Claim-Level Reliability Assessment test-time scaling), competing with models like DeepSeek V3.2 (671B) and Kimi K2.5 (1T), while also achieving a 96.1% acceptance rate (123/128 submissions) on recent unseen LeetCode weekly and biweekly contests from April–May 2026, and reaching the performance range of top-tier frontier reasoning systems including Qwen3.6 Plus and Gemini 3 Pro on verifiable reasoning benchmarks. Inference is recommended via vLLM or SGLang with temperature 1.0, top_p 0.95, and supports up to 102K output tokens, making it best suited for competitive mathematics, coding contests, and STEM reasoning tasks where clear verification signals exist, while larger general-purpose models remain preferable for broad open-domain knowledge tasks.
|
||
|
||
## Model Files
|
||
|
||
File Name | Quant Type | File Size | File Link |
|
||
|-----------|------------|-----------|-----------|
|
||
| VibeThinker-3B.BF16.gguf | BF16 | 6.18 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.BF16.gguf) |
|
||
| VibeThinker-3B.F16.gguf | F16 | 6.18 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.F16.gguf) |
|
||
| VibeThinker-3B.F32.gguf | F32 | 12.3 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.F32.gguf) |
|
||
| VibeThinker-3B.Q2_K.gguf | Q2_K | 1.27 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q2_K.gguf) |
|
||
| VibeThinker-3B.Q3_K_L.gguf | Q3_K_L | 1.71 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q3_K_L.gguf) |
|
||
| VibeThinker-3B.Q3_K_M.gguf | Q3_K_M | 1.59 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q3_K_M.gguf) |
|
||
| VibeThinker-3B.Q3_K_S.gguf | Q3_K_S | 1.45 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q3_K_S.gguf) |
|
||
| VibeThinker-3B.Q4_0.gguf | Q4_0 | 1.82 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q4_0.gguf) |
|
||
| VibeThinker-3B.Q4_K_M.gguf | Q4_K_M | 1.93 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q4_K_M.gguf) |
|
||
| VibeThinker-3B.Q4_K_S.gguf | Q4_K_S | 1.83 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q4_K_S.gguf) |
|
||
| VibeThinker-3B.Q5_0.gguf | Q5_0 | 2.17 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q5_0.gguf) |
|
||
| VibeThinker-3B.Q5_K_M.gguf | Q5_K_M | 2.22 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q5_K_M.gguf) |
|
||
| VibeThinker-3B.Q5_K_S.gguf | Q5_K_S | 2.17 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q5_K_S.gguf) |
|
||
| VibeThinker-3B.Q6_K.gguf | Q6_K | 2.54 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q6_K.gguf) |
|
||
| VibeThinker-3B.Q8_0.gguf | Q8_0 | 3.29 GB | [Download](https://huggingface.co/prithivMLmods/VibeThinker-3B-GGUF/blob/main/VibeThinker-3B.Q8_0.gguf) |
|
||
|
||
## llama.cpp
|
||
|
||
LLM inference in C/C++ — https://github.com/ggml-org/llama.cpp |