This illustrates a widespread problem affecting large language models (LLMs): even when an English-language version passes a safety test, it can still hallucinate dangerous misinformation in other ...

This release is good for developers building long-context applications, real-time reasoning agents, or those seeking to reduce GPU costs in high-volume production environments.

What makes this particularly dangerous in enterprise and production contexts is not just that the model gets it wrong, but ...

OpenAI Group PBC and Mistral AI SAS today introduced new artificial intelligence models optimized for cost-sensitive use cases. OpenAI is rolling out two algorithms called GPT-5.4 mini and GPT 5.4 ...

As large language models (LLMs) gain momentum worldwide, there’s a growing need for reliable ways to measure their performance. Benchmarks that evaluate LLM outputs allow developers to track ...

Nvidia's KV Cache Transform Coding (KVTC) compresses LLM key-value cache by 20x without model changes, cutting GPU memory costs and time-to-first-token by up to 8x for multi-turn AI applications.

MIT study finds cross-model uncertainty measurement outperforms traditional methods in spotting unreliable AI predictions ...

Mark Stevenson has previously received funding from Google. The arrival of AI systems called large language models (LLMs), like OpenAI’s ChatGPT chatbot, has been heralded as the start of a new ...

As great as generative AI looks, researchers at Harvard, MIT, the University of Chicago, and Cornell concluded that LLMs are not as reliable as we believe. Even a big company like Nintendo did not ...