Chinese AI lab DeepSeek has released an open version of DeepSeek-R1, its so-called reasoning model, that it claims performs as well as OpenAI's o1 on certain AI benchmarks. R1 is available from the AI ...

DeepSeek-R1 demonstrates strong performance across multiple educational and reasoning benchmarks. It achieves 79.8% Pass@1 on AIME 2024, slightly surpassing OpenAI-o1-1217. On MATH-500, it reaches ...

OpenAI o1 leads in reasoning tasks due to its ability to “think” before responding, while DeepSeek R1 offers competitive performance at a significantly lower cost. Llama 3.2 shines in multimodal use ...

The abrupt appearance of DeepSeek’s R1 advanced reasoning model at the start of the year was akin to the “shot heard ‘round the world” in AI circles. Major tech companies had spent recent years ...

Researchers have introduced Light-R1-32B, a new open-source AI model optimized to solve advanced math problems. It is now available on Hugging Face under a permissive Apache 2.0 license — free for ...