Particle.news
Download on the App Store

OpenAI Releases GPT‑5.4 Mini and Nano for Low‑Latency, Low‑Cost AI Tasks

Purpose‑built for low‑latency subagents, the models pair near‑flagship accuracy with pricing that makes multi‑model pipelines practical.

Overview

  • GPT‑5.4 mini is live in ChatGPT for Free and Go users via the Thinking option and in the API and Codex, while GPT‑5.4 nano launches API‑only for developer use.
  • OpenAI lists GPT‑5.4 mini at $0.75 per million input tokens and $4.50 per million output tokens with a 400k context window; GPT‑5.4 nano costs $0.20 and $1.25 respectively.
  • Benchmarks reported by OpenAI show mini nearing the flagship on coding and computer‑use tests (e.g., 54.38% on SWE‑Bench Pro and 72.13% on OSWorld‑Verified) while running more than 2x faster than GPT‑5 mini.
  • Mini serves as a cost‑saving executor in Codex, using about 30% of a standard GPT‑5.4 quota, as OpenAI promotes hybrid architectures where a planning model delegates to cheaper subagents.
  • Early tests from customers and third parties, including Hebbia, Notion, and Perplexity, cite strong task performance for mini and responsiveness and efficiency gains for nano in live workflows.