TLDR AI · RESEARCH & INNOVATION
DEEPSEEK PROVER
Summary from the TLDR AI newsletter
DeepSeek has produced an incredibly powerful math model that was trained via process supervision and dramatically outperforms larger models on various MATH benchmarks.