Scaling01 on LLM Pricing: Linear Cost Scaling, Margin Drives Price
scaling01 · x · 2026-08-31
In a discussion on LLM pricing models, the author argues that linear pricing is a valid first-order approximation for frontier models. Key points:
- Cost & Params: Cost depends mostly on active params (3-5% of total), scaling linearly with total params.
- Scaling Regime: Scaling laws hold better for large frontier models designed for compute-constrained regimes than for smaller ones.
- Pricing Power & Margins: Price = Cost × Markup. Frontier models maintain high margins due to competitiveness, whereas most open-weight models lack pricing power and see margins shrink as size increases (with rare exceptions like Kimi-K3 and GLM-5.3).
The author argues that listing random open models without pricing power does not refute this analysis.
Related event: Debate: Does Parameter Count Determine Frontier Model Pricing?(9 posts)→
More from Infra
- Trading 2 RTX 5090s for 4 R9700s: does doubled VRAM pay off? — poofph · 2026-10-16
- Running PolyStrata + Cline coding agent on a $0-fancy laptop at ~90 tok/s — rpwoerk · 2026-10-16
- TensorFold 1.0.12 runs 120B vision model on 1-2 DGX Sparks, adds Mac image input — HankYeomans · 2026-10-16
- vLLM brings frontier model local deployment to NVIDIA RTX Spark PCs — vllm_project · 2026-10-16
- NVIDIA signals $2B in new European AI deals at GTC Berlin, totaling ~$10B in a year — CurieuxExplorer · 2026-10-16
- Making Your Data Ready for Agentic AI: trust first, then semantics, then writes — bibryam · 2026-10-16