Shareable benchmark table
GLM-5.3 arm conditions and budgets
The three frozen conditions of the parent pilot, reused unchanged (same turns, tools, and wall caps) over the same three tasks.
Table 16 · Frozen arm protocol
GLM-5.3 arm conditions and budgets
| Condition | Agent access | Turn / tool budget | Wall cap |
|---|---|---|---|
| Tool-assisted | Bounded search/read + forge build/test feedback, then one submitted proof | 12 turns / 10 tools | 1500 s |
| Static (no tools) | Full source context (target + provided test harness); no tool feedback | 3 turns / 0 tools | 1200 s |
| Budgeted triage | A strict budget of four inspect operations, then one submission | 6 turns / 4 tools | 1200 s |
Wall time runs begin→final and includes GLM-5.3 + ZCode inference between turns. The agent interacts with the task workspace only through the driver, which enforces the parent action policy; the fixed revision is never mounted.
