---

license: mit

base_model:

• WeiboAI/VibeThinker-1.5B

pipeline_tag: text-generation

tags:

• math
• code
• reasoning
• gpqa
• instruction-following
• gguf

---

constructai/VibeThinker-1.5B-GGUF

This is a quantized version of the original VibeThinker-1.5B , converted to the GGUF format for efficient CPU/GPU inference with llama.cpp, Ollama, or any GGUF‑compatible runner.

---

Original Model

• Author(s): WeiboAI
• Source: VibeThinker-1.5B
• Original License: MIT

---

Available Quantizations

Choose the quantization that fits your needs:

| Quantization | File Size |

|--------------|-----------|

| UD-IQ1_S | 437 MB |

| UD-IQ1_M | 464 MB |

| UD-IQ2_XXS | 511 MB |

| Q2_K | 676 MB |

| UD-IQ2_M | 601 MB |

| UD-Q2_K_XL | 676 MB |

| UD-IQ3_XXS | 669 MB |

| Q3_K_S | 761 MB |

| UD-IQ3_S | 762 MB |

| Q3_K_M | 824 MB |

| UD-Q3_K_M | 824 MB |

| UD-Q3_K_XL | 880 MB |

| UD-IQ4_XS | 896 MB |

| Q4_K_S | 940 MB |

| UD-IQ4_NL | 936 MB |

| Q4_K_M | 986 MB |

| UD-Q4_K_XL | 986 MB |

| Q5_K_S | 1.1 GB |

| UD-Q5_K_S | 1.1 GB |

| Q5_K_M | 1.13 GB |

| UD-Q5_K_M | 1.13 GB |

| UD-Q5_K_XL | 1.13 GB |

| Q6_K | 1.27 GB |

| UD-Q6_K | 1.27 GB |

| UD-Q6_K_XL | 1.27 GB |

| Q8_0 | 1.65 GB |

| UD-Q8_K_XL | 1.65 GB |

| F16 | 3.09 GB |

For a 1.5B‑parameter model, even the larger files are quite manageable. Here’s what I recommend: F16 (3.09 GB) or Q8_0 (1.65 GB).

The other quants are also usable!

---

Usage

With ollama

ollama run hf.co/constructai/VibeThinker-1.5B-GGUF:F16

---

With llama.cpp

llama-server -hf constructai/VibeThinker-1.5B-GGUF:VibeThinker-1.5B-GGUF-F16.gguf

or

llama-cli -hf constructai/VibeThinker-1.5B-GGUF:VibeThinker-1.5B-GGUF-F16.gguf

---

With LM Studio

lms get constructai/VibeThinker-1.5B-GGUF@F16

---