Artificial Analysis: frontier-level intelligence gets 50% cheaper within about six weeks of each threshold
ArtificialAnlys · x · 2026-10-21
Artificial Analysis launched a "cost of intelligence over time" tracker. Key finding: for each Intelligence Index threshold, cost per task fell at least 50% within ever-shortening windows from the first measured model—197 days for 10, 169 for 20, 65 for 30, 42 for 40, and just 41 days for 50, meaning reaching 40 or 50 became at least 50% cheaper within roughly six weeks. The page also tracks quarterly capex by Microsoft, Google, Meta, Amazon and Oracle, and frontier model intelligence over time (Index v4.3.2, 10 evaluations, 16 of 60 model creators).
More from Infra
- PyTorch details KTIR and Dataflow Scheduler behind IBM's Spyre accelerator backend — PyTorch · 2026-10-21
- Indie dev replaces $1k OpenAI bill with a YAML pipeline on spot GPUs and vLLM — cmm324 · 2026-10-21
- DeepSeek open-sources DeepInfer, the vLLM-based engine behind V4.1-Flash — The Decoder · 2026-10-21
- Robotics startup launches Deployments to serve fine-tuned HF checkpoints to robots in real time — arnie_hacker · 2026-10-21
- ByteDance, Alibaba, Tencent order 3M+ Nvidia RTX PRO 5500 cards worth $35B, Reuters reports — SumitGup · 2026-10-21
- TorchTPU demo: unified vLLM & SGLang backend for LLM serving on TPUs — Rob_Mulla · 2026-10-21