Generated: 2026-05-02T07:38:40+00:00
Suite: qorx-self
Target: .
Qorx version: qorx 1.0.4
Git commit: 78469bd
| Metric | Value |
|---|---|
| Indexed local tokens | 219838 |
| Session visible tokens | 73 |
| Session reduction | 3011.48x |
| Pack used tokens | 484 |
| Pack reduction | 454.21x |
| Squeeze used tokens | 419 |
| Squeeze reduction | 524.67x |
| Bench average reduction | 451.99x |
| Strict task pass rate | 100.0% |
| Expected refusal pass rate | 100.0% |
| Agent provider calls | 0 |
| Question | Expected | Actual | Pass | Evidence | Used tokens |
|---|---|---|---|---|---|
| context fault proof pages resolver boundary | supported | supported | yes | 3 | 380 |
| galactic banana escrow treaty | not_found | not_found | yes | 0 | 8 |
| Query | Used tokens | Omitted tokens | Reduction | Quarks |
|---|---|---|---|---|
| context fault proof pages resolver boundary | 484 | 219354 | 454.21x | 3 |
| qorx carriers .qorx .qorxb qorx handle | 450 | 219388 | 488.53x | 3 |
| strict answer refusal unsupported claims | 532 | 219306 | 413.23x | 4 |
This benchmark uses Qorx local accounting only. Token counts are deterministic
ceil(chars / 4) estimates unless the runtime reports another estimator. The
report does not claim provider invoice savings, production throughput, or
downstream model answer quality.
To reproduce:
python scripts/run-benchmark.py --target . --suite qorx-self --budget-tokens 600 --squeeze-budget-tokens 450 --query "context fault proof pages resolver boundary" --query "qorx carriers .qorx .qorxb qorx handle" --query "strict answer refusal unsupported claims" --supported-question "context fault proof pages resolver boundary" --unsupported-question "galactic banana escrow treaty" --agent-objective "prove context fault proof pages resolver boundary" --output-json docs/benchmarks/2026-05-02-qorx-self.json --output-md docs/benchmarks/2026-05-02-qorx-self.md