3 articles
The new Mercury 2 AI model has achieved a record 78% success rate on the Pinch Bench benchmark for open-source AI agents. It also boasts a 1.7-second latency and significantly lower costs compared to leading models, making AI agents more viable.
A new AI model called Mercury 2 uses a novel 'diffusion' approach to generate text much faster than current models. This technology promises to speed up AI applications by over five times, making them more practical for real-world use.
Inception Labs has unveiled Mercury 2, the world's first Diffusion Large Language Model (DLM) capable of reasoning. This new architecture significantly outperforms traditional Transformer models in speed and accuracy, potentially reducing AI hallucinations and enabling near-instantaneous applications.
