Chinese reasoning models match the best on 4 of 5 benchmarks at 40x lower cost

1 points | by moralec 5 hours ago

5 comments