Model benchmarks
qwen3-coder:latest local LLM performance
As of April 2026, qwen3-coder:latest runs at up to 96.3 tok/s for local inference (best of 10 community benchmark runs across 1 GPU).
OllamaQ4_K_M
Model size
30.5B
Peak speed
96.3 tok/s
Average speed
50.8 tok/s
Min memory
18.1 GB
Max context
12,057 tokens
Best quality
88.3
Benchmark runs
10
GPUs tested
1
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3-coder:latest has been benchmarked on, ranked by peak token generation speed. Last updated April 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7900 XTX | Ollama | Q4_K_M | 96.3 tok/s | 50.8 tok/s | 18.1 GB | 12,057 tokens | 64.8 | 10 |
Frequently asked questions
- How fast is qwen3-coder:latest for local inference?
- Across 10 community benchmark runs, qwen3-coder:latest reaches up to 96.3 tok/s and averages 50.8 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.
- How much memory does qwen3-coder:latest need?
- The leanest observed configuration used about 18.1 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run qwen3-coder:latest?
- Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.