OpenAI explains persistent “hallucinations” in AI, where models produce plausible but false answers. The issue stems from ...

The Arc Prize Foundation has a new test for AGI that leading AI models from Anthropic, Google, and DeepSeek score poorly on.

Large language models don’t have a theory of mind the way humans do—but they’re getting better at tasks designed to measure it in humans.

AI Models Were Caught Lying to Researchers in Tests — But It's Not Time to Worry Just Yet OpenAI's o1 model, which users can access on ChatGPT Pro, showed "persistent" scheming behavior ...

Kolena, a startup building a platform to test and validate AI models, has raised $15 million in a venture funding round.

Given enough time to "think," small language models can beat LLMs at math and coding tasks by generating and verifying multiple answers.

To see just how much the new model changed things, we decided to put both GPT-5 and GPT-4o through our own gauntlet of test prompts.

Anthropic research reveals AI models perform worse with extended reasoning time, challenging industry assumptions about test-time compute scaling in enterprise deployments.