Reaction
DeepSeek is said to fall short of an earlier, smaller Dots3-preview on performance and cost
A user calls the comparison a strong vindication of the Dots3 team’s reinforcement-learning scheme, while describing Dots3 as weaker across the board.
TLDR
A user says DeepSeek did not match an earlier, smaller Dots3-preview on either performance or cost. They describe Dots3 as weaker across the board, but see the comparison as a strong vindication of the team’s reinforcement-learning scheme and urge more attention for the team.
Combined views
3.2K
1 Source, first seen ago