The original DeepSeek R1 versus newer DeepSeek models
A user says newer DeepSeek models feel more “sane” and likens many to a less uptight Claude, but calls the original R1 “something else.”
TLDR
A user thinks DeepSeek’s personality training is less harsh than Anthropic’s training for Claude, and says that accounts for some differences they notice. They recommend trying the original DeepSeek R1 at least once, saying it remains available because it is open source.
Combined views
10.1K
2 Sources, first seen 4h ago
likes