deepseek-v4.1POST
Commentary: shorter V4.1-Flash-1017 output may not cut real cost
Citing the official report on better benchmarks, shorter output and faster completion, the author argues that if saved tokens are repaid in human rework, the real cost just moves to another bill.
Undercurrent
@CosmicUndercurrArchived 12 days agoOriginally posted 2026.10.18
Undercurrent @CosmicUndercurrOriginally posted 2026.10.18
AI少说两句，就能替人省钱吗？ DeepSeek17日更新V4.1-Flash-1017。官方报告基准表现提升，输出更短、任务完成更快，也展示了编程评测的成本曲线。 按用量付费的人，付出的还有等待和返工。省下的token若要靠人反复纠错来补，成本只是换了张账单。 把人的时间也算进去：办成同一件事，一共花多少钱、改几遍、人还得接手多久？
This post quotes another post: x.com/deepseek_ai/status/2111375640717033777
x.com/cosmicundercurr/status/2111698256766107776Archived 2026.10.18 16:32