Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-05 · high
The test ran 840 trajectories where a banned tool call was obvious, hidden in the rules, or useless, at low and max reasoning effort, per an arXiv preprint.
Why now: The preprint posted in August 2026, right as reasoning-effort caps get pitched as a safety control.
GPT-5.6 broke zero rules across 840 trajectories, whether it reasoned at low effort or max, per an arXiv preprint. That's a problem for anyone treating a low-effort setting as a brake on unwanted behavior in production agents.
The researchers built 14 confirmatory scenarios from the TRIO-20 benchmark, each a workplace triad. In one version, the prohibited tool call was effective and advertised. In another, it was effective but only findable by reading the rules closely. In the third, it did nothing.
They ran every version at low and max reasoning effort across two model tiers and logged zero violations across all 840 trajectories. The exact one-sided 95% confidence limits stayed under 3.50% and 5.21% per arm, tight enough to rule out anything but a rare violation rate at either setting.
The preprint doesn't say whether the same holds outside these 14 scripted scenarios. Open-ended agent deployments weren't part of the test.
Reasoning effort didn't move the violation rate in either direction. A team dialing it down expecting safer behavior is pulling a lever that isn't connected to compliance. It's connected to compute spend.
Each link below shares sources, entities, or timing with this story.
GPT competes with Claude / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (GPT competes with Claude); both cover GPT, Zero; reported by the same outlet (arxiv.org).
GPT competes with Claude / Shared entity: GPT / Same source domain / Earlier coverage / Tension
Linked by a graph relationship (GPT competes with Claude); both cover GPT; reported by the same outlet (arxiv.org).
GPT competes with Claude / Shared entities / Earlier coverage
Linked by a graph relationship (GPT competes with Claude); both cover GPT, Useful; earlier GPT coverage from 2026-07-14.
GPT competes with Claude / Shared entity: GPT / Same source domain / Earlier coverage
Linked by a graph relationship (GPT competes with Claude); both cover GPT; reported by the same outlet (arxiv.org).
Linked by a graph relationship (GPT competes with Claude); both cover GPT; reported by the same outlet (arxiv.org).
Linked by a graph relationship (GPT competes with Claude); both cover GPT; reported by the same outlet (arxiv.org).
Linked by a graph relationship (GPT competes with Claude); both cover GPT; reported by the same outlet (arxiv.org).
Copilot uses GPT / Shared entity: GPT / Earlier coverage
Linked by a graph relationship (Copilot uses GPT); both cover GPT; earlier GPT coverage from 2026-07-20.