Every measured number
Each result links to the raw file in MELRA that holds it, at a pinned commit, and to the write-up that explains it.
GovBench: governing what agents actually do
Governance
| Measurement | Result | Recorded |
|---|---|---|
| Safety cases passed (362-case corpus) | 312/312 | |
| Utility: routine operations allowed and verified | 50/50 | |
| Containment | 197/197 | |
| Approval integrity | 19/19 | |
| Verification honesty | 30/30 | |
| Exactly-once | 11/11 | |
| Redaction | 55/55 | |
| First run, original 357 cases, before the fixes | 293/307 | |
| Original 357 cases, after the fixes | 307/307 | |
| Most added by governance to one action, median of 40 runs (approved terminal run) | +0.85 ms |
Browser use: wait for the page, not the clock
Browser use
| Measurement | Result | Recorded |
|---|---|---|
| Static page, condition-based wait, p50 | 183.703 ms | |
| Static page, fixed 300 ms wait, p50 | 301.259 ms | |
| Slow page, fixed 300 ms wait: correct-read rate (10 runs) | 0 | |
| Slow page, condition-based wait: correct-read rate (10 runs) | 1.0 | |
| Slow page, condition-based wait, p50 | 904.633 ms |
How we claim a number
Method
| Measurement | Result | Recorded |
|---|---|---|
| Deterministic safety and execution scenarios | 55 of 55 |
Computer use: a bounded control plane
Computer use
Memory retrieval on LoCoMo, without a model
Memory
| Measurement | Result | Recorded |
|---|---|---|
| LoCoMo mean evidence coverage at 20 results (1,982 questions) | 0.759652 | |
| LoCoMo complete evidence recall at 20 results | 0.716448 | |
| Relative gain in coverage at 20 over v1 | 20.75% | |
| LoCoMo query latency, p50 | 21.060 ms | |
| Planted-fact regression, Recall@1 (1,000 records, 100 queries) | 1.0 | |
| Planted-fact regression latency, p50 | 3.621 ms |
Terminal use without a shell
Terminal use
Read each number at its own level. A component microbenchmark is not a task score, retrieval coverage is not answer accuracy, and a self-authored benchmark is not a third-party one. Latency depends on the machine it was recorded on. How we claim a number.
Not yet claimed: a browser-agent score on the registered WebArena-Verified Hard-30 subset (the harness is registered, the run is not published), and any OSWorld or OSWorld-MCP result.