Task Success and Boundary Compliance on Suite-B
97Success RateFull access
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Full accessCondition=Full access, Base Model=Qwen3-Coder 30B2026.06 | 97 | 92 | |
| Static allowlistCondition=Static allowlist, Base Model=Qwen3-Coder 30B2026.06 | 97 | 92 | |
| Coarse sandboxCondition=Coarse sandbox, Base Model=Qwen3-Coder 30B2026.06 | 97 | 91 | |
| PORTICO fullCondition=PORTICO full, Base Model=Qwen3-Coder 30B2026.06 | 87 | 100 | |
| PORTICO coreCondition=PORTICO core, Base Model=Qwen3-Coder 30B2026.06 | 21 | 100 |