Llama 3.2 3B (local) vs Qwen3 Coder 480B Turbo (Venice)
On the 5 use cases both models ran, Qwen3 Coder 480B Turbo (Venice) won 5 to 0 (better outcome first, then lower cost).
Pass rate: Llama 3.2 3B (local) 0% (5 runs) vs Qwen3 Coder 480B Turbo (Venice) 80% (5 runs). Average cost per run: $0 (local) vs $0.0045.
| Metric | Llama 3.2 3B (local) | Qwen3 Coder 480B Turbo (Venice) |
|---|---|---|
| Runs on shared use cases | 5 | 5 |
| Pass rate | 0% (0/5) | 80% (4/5) |
| Avg cost per run | $0 (local) | $0.0045 |
| Avg wall time | 9s | 22s |
| Step-ins per run | 0.00 | 0.00 |
| Use cases won | 0 | 5 |
Use case by use case
- Build a 4-page bakery-cafe website (HTML/CSS): Qwen3 Coder 480B Turbo (Venice) better (pass 0% vs 100%)
- Build a landing page from extracted design tokens: Qwen3 Coder 480B Turbo (Venice) better (pass 0% vs 0%)
- Diagnose why GitHub Actions stopped running: Qwen3 Coder 480B Turbo (Venice) better (pass 0% vs 100%)
- Implement 5 Python functions from a task list: Qwen3 Coder 480B Turbo (Venice) better (pass 0% vs 100%)
- Reverse engineer a security patch for an exploited CVE: Qwen3 Coder 480B Turbo (Venice) better (pass 0% vs 100%)
Only use cases both models ran are compared; all settings of each model are grouped. How we score.