DeepSeek V4.1F serving is far cheaper than rivals — and DS deliberately overprices its tokens
teortaxesTex · x · 2026-10-16
Citing tphuang's observation that Tencent and Zhipu both jammed up almost immediately hosting DeepSeek V4.1F — leaving them entirely reliant on DeepSeek's own servers — teortaxesTex argues the root cause is economics: serving V4.1F is much cheaper than any competing model, and the stack is better too. The gap is more drastic than API prices suggest, because DeepSeek doesn't want to dedicate much capacity to inference and deliberately overprices its tokens to curb demand. He quips: if Zhipu can't handle it with a 100k-GPU cluster, how is DeepSeek managing it?
More from Infra
- Bonsai 2 2B hits 120 tok/s on iPhone 18 Pro, outpaces Qwen 3.5 2B at half the memory — cephaloform · 2026-10-16
- Samsung cuts Q4 phone output as Nittobo shifts E-glass to AI-chip T-glass — zephyr_z9 · 2026-10-16
- White House AI chief discusses Empire AI, New York's shared compute infrastructure — mkratsios47 · 2026-10-16
- llama.cpp merges Muse Spark Open (816B-A42B) text, vision and audio support in 4 hours — jacek2023 · 2026-10-16
- FT warns of 'Darwin moment' for AI borrowers as CoreWeave bonds sink to distress levels — firstadopter · 2026-10-16
- Phone as interface, home servers as the lab: a two-workflow local AI setup — EAccelerate_42 · 2026-10-16