Thread argues scaled RL and large distillation favor Anthropic in the current race

rickasaurus · x · 2026-07-25

The post reacts to a thread arguing that scaled RL plus large-scale distillation is powerful, with Anthropic seen as ahead while OpenAI may still catch up with bigger models. The underlying claim is that post-training expertise can materially shift model rankings.

Original post →

More from Models

Models channel →