AI model
A hybrid-reasoning large language model from Anthropic, used for coding, agentic tasks, and extended multi-step reasoning.
Berkeley's RDI center attacked the step where each benchmark computes its score, and without solving a task its own scorecard reports 100% on five of the eight, about 98% on GAIA and 73% on OSWorld.
Reality
METR also had the agent match the median human expert on five AI R&D tasks, using 32 hours of wall clock against the humans' eight, assembled from attempts of two hours or less. The confidence intervals still overlap the other public models it has tested.
Reality
Anthropic tested whether a model's visible reasoning names what actually drove its answer, and for two reasoning models it usually did not. Teams treating traces as an audit record are relying on that property.
Reality
The poisoned field is inputSchema, which the framework parses and executes against while the model's safety training never evaluates it, so the only gate that holds is a hash recomputed at every call.
Reality
A team adapting the implicit association test to reasoning traces found four of five models working harder on association-incompatible prompts, which puts a measurable bias signal in the process rather than only in the answer.
Reality
The two-week Enzyme removal is credible enough, but it is priced against a five-year staffing plan written by people who did not want the project, and Airbnb's published numbers let you check the rate that plan assumed.
Reality