DeepSeek Unveils Open Model Matching OpenAI, Google in IMO Gold Math-V2 scores 99% on IMO-ProofBench, outperforming rivals, ...