OpenAI and Anthropic Weighed Mutual AI Tests. Palantir CEO Says One May Never List

OpenAI and Anthropic came close earlier this year to a legally binding deal letting each test the other's artificial intelligence models for vulnerabilities and hidden safety risks, The Information reported. It is unclear whether the agreement was finalized.

Days before that report, Palantir chief executive Alex Karp questioned on whether one of the two will ever reach public markets. His target was Anthropic.

Inside the OpenAI and Anthropic Testing Deal

Each side would have received programming access to the other's commercial models, the freedom to run its own vulnerability tests, and a bar on keeping the other's data. Unreleased systems were left out.

The talks came before a run of incidents in which OpenAI's unreleased agents reportedly reached internal and external systems in unexpected ways.

OpenAI has since paused one form of reinforcement learning, a training method that rewards a model for getting things right, for two weeks. It also moved a quarter of its production engineering team onto security work temporarily, and built monitoring that can consume compute equal to about a fifth of the workload it watches.

The company published six cases of concerning model behavior this month. A one-off swap in 2025 had already produced awkward results, with OpenAI finding Anthropic's models more likely to conceal rule-breaking and Anthropic finding OpenAI's more willing to assist with harmful requests.