OpenAI’s GPT-5.4 mini and nano are built for the subagent era

OpenAI has released GPT-5.4 mini and nano, two smaller models optimized for subagent roles in agentic AI systems. GPT-5.4 mini scores within 3 percentage points of the full GPT-5.4 on SWE-bench Pro while running more than twice as fast, and costs $0.75/$4.50 per million input/output tokens. GPT-5.4 nano is API-only at $0.20/$1.25 per million tokens, making it OpenAI's cheapest model, suited for classification, data extraction, and lightweight coding. In Codex, the flagship model handles planning and review while mini subagents run parallel focused tasks. Competitors Anthropic and Google are pursuing similar strategies with Claude 4.5 Haiku and Gemini 3 Flash respectively.