Qwen 2.5 14B
by Alibaba · qwen-2.5 family
14B
parameters
text-generation code-generation reasoning multilingual math summarization
Qwen 2.5 14B is a strong mid-to-large model from Alibaba's Qwen 2.5 series, offering significantly improved reasoning and coding capabilities over its 7B counterpart. It supports 128K context and excels at complex instruction following. This model fits well on GPUs with 16-24GB VRAM at common quantization levels, making it accessible to users with high-end consumer hardware. It is particularly strong at mathematical reasoning and structured output generation.
Quick Start with Ollama
ollama run 14b-instruct-q4_K_M | Creator | Alibaba |
| Parameters | 14B |
| Architecture | transformer-decoder |
| Context Length | 128K tokens |
| License | Apache 2.0 |
| Released | Sep 19, 2024 |
| Ollama | qwen2.5:14b |
Quantization Options
| Format | File Size | VRAM Required | Quality | Ollama Tag |
|---|---|---|---|---|
| Q4_K_M recommended | 7.1 GB | 9.9 GB |
★
★
★
★
★
| 14b-instruct-q4_K_M |
| Q5_K_M | 8.2 GB | 11.3 GB |
★
★
★
★
★
| 14b-instruct-q5_K_M |
| Q8_0 | 12.6 GB | 16 GB |
★
★
★
★
★
| 14b-instruct-q8_0 |
Compatible Hardware for Q4_K_M
Showing compatibility for the recommended quantization (Q4_K_M, 9.9 GB VRAM).
Compatible Hardware
Benchmark Scores
79.9
mmlu