Leaving r1 at 15.8% percent with $0.06/task without scaling to medium and high c...

		georgiivanov on Jan 30, 2025 \| parent \| context \| favorite \| on: An analysis of DeepSeek's R1-Zero and R1 Leaving r1 at 15.8% percent with $0.06/task without scaling to medium and high compute is a bit misleading imo. The whole point of DeepSeek is the efficiency.