LLM
GA
Open source
Qwen3.8-27B
Best open general-purpose model on a single 24 GB GPU
Release2026-09
Context262K
Max output65,536 tokens
Parameters27B
Local / VRAM18–24 GB
LicenseOpen source
Overview
Qwen3.8-27B is a model in the LLM category from Alibaba (Qwen). It was released in 2026-09. The vendor is based in China. Best for: Best open general-purpose model on a single 24 GB GPU.
Hardware: Q4 ~16-18 GB; comfortable on a 24 GB GPU.
Strengths
- Apache-2.0
- 262K native / 1M YaRN
- Visual input
Limits
- Behind frontier closed models
- Slows down at long context
Specifications
| Vendor | Alibaba (Qwen) |
|---|---|
| Category | LLM |
| Release | 2026-09 |
| Status | GA |
| License | Open source |
| License name | Apache-2.0 |
| Commercial use | Allowed |
| Parameters | 27B |
| Architecture | dense transformer |
| Context window | 262K |
| Max output tokens | 65,536 |
| Input modalities | Text, Image |
| Output modalities | Text |
| Prompt types | Text |
| Negative prompts | No |
| Max outputs | 1 |
| Local deployment | Yes |
| VRAM | 18–24 GB |
| System RAM | 32–64 GB |
| API access | Yes |
| Pricing | Free weights |
| Confidence | High |
| Vendor country | China |
| Vendor type | Cloud provider |
Pricing & hardware
Free weights
This model can be self-hosted or run locally. Approximate GPU memory: 18–24 GB. Approximate system RAM: 32–64 GB. API access is available.