Particle.news

AI Labs Slash Prices as Phones, Chips and EVs Rush to Run Smarter Agents

Steep API cuts from OpenAI and Anthropic plus new consumer and auto hardware are lowering per‑use AI costs and driving much higher demand for inference and on‑device intelligence.

Overview

  • OpenAI introduced GPT‑6Sol and GPT‑6Luna with far lower API rates and Anthropic released Claude Opus 5.5 at roughly 40% lower runtime cost, forcing a fresh round of per‑token price declines across the market.
  • Anthropic says Opus 5.5 is faster, expands multimodal and agent‑style workflows, and is being used inside the company to automate portions of its R&D work.
  • Meta’s Muse has moved beyond chat into task‑doing agents and recorded rapid early uptake in the U.S. and Canada with downloads and daily active users that outpaced ChatGPT in the same launch window.
  • Device and chip vendors answered with new hardware: Qualcomm unveiled a second‑generation Snapdragon Audio Platform and new Snapdragon mobile platforms, Xiaomi announced the Xiaomi 18 Pro series with a new Snapdragon flagship, and Huawei and BYD began deliveries and launches of HarmonyOS‑stacked and long‑range EVs respectively.
  • Lower model prices are already shifting the economics of AI by making high‑frequency, long‑context and agent workloads far cheaper to run, which will raise demand for data‑center and edge compute and increase electricity use while squeezing upstream hardware margins.