ai/qwen3-coder
Verified Publisher
Qwen3-Coder is Qwen’s new series of coding agent models.
Model
16
50K+
ai/qwen3-coder repository overview
Qwen3‑Coder‑30B‑A3B‑Instruct
GGUF version by Unsloth
Open‑source agentic coding model optimized for long‑context, instruction‑following code generation and tooling.
Intended uses
Lightweight yet powerful coding assistant for repository‑scale tasks:
- Agentic coding workflows: Automate multi‑step code tasks with platform integrations (e.g. Qwen Code, CLINE).
- Browser‑use scenarios: Drive agentic browser interactions for search, scraping, or UI automation.
- Large code context comprehension: Handle repository‑scale files with native support up to 256K token context (extendable to ~1M via Yarn).
Characteristics
| Attribute | Details |
|---|---|
| Provider | Qwen / Alibaba |
| Architecture | MoE (Mixture of Experts, 30.5B total with ~3.3B active, 128 experts, 8 active) |
| Cutoff date | July 2025 (model released August 2025) |
| Languages | Multilingual; over 100 spoken and programming languages |
| Tool calling | Yes |
| Input modalities | Text (code + natural language) |
| Output modalities | Text (code + natural language) |
| License | Apache‑2.0 |
Available model variants
| Model variant | Parameters | Quantization | Context window | VRAM¹ | Size |
|---|---|---|---|---|---|
ai/qwen3-coder:30B-A3B-UD-Q4_K_XLai/qwen3-coder:latest | 30B-A3B | MOSTLY_Q4_K_M | 262K tokens | 17.52 GiB | 16.45 GB |
¹: VRAM estimated based on model characteristics.
latest→30B-A3B-UD-Q4_K_XL
Use this AI model with Docker Model Runner
First, pull the model:
docker model pull ai/qwen3-coder:30B-A3B-UD-Q4_K_XL
Then run the model:
docker model run ai/qwen3-coder:30B-A3B-UD-Q4_K_XL
For more information on Docker Model Runner, explore the documentation.
Links
Tag summary
Content type
Model
Digest
sha256:8fdf3f084…
Size
16.5 GB
Last updated
2 days ago
docker model pull ai/qwen3-coder:30BThis week's pulls
Pulls:
1,258
Last week