Comment by IanCal

> OpenAI / Anthropic models have largely stopped advancing

Have they? That seems like quite a claim given the last 6 months, particularly for cybersecurity.