OpenAI Launches GPT-5.4 Mini and Nano for Fast, Low-Cost Agentic Workflows

Mar 19, 2026 · Updated Apr 25, 2026

GPT-5.4 mini and GPT-5.4 nano are OpenAI's new small model tier, bringing GPT-5.4 capability to faster, cheaper deployments. Mini runs more than 2x faster than GPT-5 mini, scores 54.4% on SWE-Bench Pro and 72.1% on OSWorld-Verified, and carries a 400k context window at $0.75 per 1M input tokens. Nano targets classification, data extraction, ranking, and simpler coding subtasks at $0.20 per 1M input tokens, API-only.

This signals OpenAI's push toward composable multi-agent systems — a larger model handles planning while mini or nano subagents run narrower tasks in parallel at a fraction of the cost. In Codex, mini uses only 30% of the GPT-5.4 quota, giving agentic coding workflows a direct cost lever.

Swap mini into your Codex subagents for codebase search, file review, or document processing — faster responses at one-third the quota cost. Mini is also live in ChatGPT for Free and Go users via the Thinking feature.

View the full update on openai.com

OpenAI Developers

@OpenAIDevsMar 17

We’re introducing GPT-5.4 mini and nano, our most capable small models yet. GPT-5.4 mini is more than 2x faster than GPT-5 mini. Optimized for coding, computer use, multimodal understanding, and subagents. For lighter-weight tasks, GPT-5.4 nano is our smallest and cheapest version of GPT-5.4. https://t.co/6E6cP6saWT

613

View on X