GPT-5.4 mini

LLM Provider / Model · Coding, computer use and subagent orchestration on a budget

At a glance

Input price	$0.75
Output price	$4.50
Cache price	$0.075
Price tier	Budget
Context window	400K
Max output	128K
Context tier	128K-500K
Speed tier	Fast
Latency	fast
Knowledge cutoff	Aug 2025
Modality	Multimodal (vision)
Model ID	gpt-5.4-mini
Provider	OpenAI
Benchmark	Budget tier with computer use and subagent support
Released	Mar 2026
Hosting	Closed/API
Capabilities	Reasoning effort control (none-xhigh), Computer use, Subagents, Vision, Tool use, Structured output

Reasoning effort control (none-xhigh), Computer use, Subagents, Vision, Tool use, Structured output

Coding, computer use and subagent orchestration on a budget

Trae's built-in model menu may not include GPT-5.4 mini — Trae primarily ships with Claude/GPT/Gemini/DeepSeek.
Kiro's spec mode is tuned for Claude models — GPT-5.4 mini may work but spec-driven flows are best with Claude.
Gemini CLI is locked to Google Gemini models — your LLM is GPT-5.4 mini.
Claude Code only works with Claude models — your LLM is GPT-5.4 mini.
Devin uses Cognition's proprietary model internally — your selected GPT-5.4 mini will be ignored.
OpenAI Codex (Cloud) runs GPT-5.1 Codex internally — your selected GPT-5.4 mini will be ignored.
Jules runs Gemini models internally — your selected GPT-5.4 mini will be ignored.
The Claude Code SDK is tuned for Claude models — GPT-5.4 mini may work via the model adapter but loses Claude-specific features (thinking, prompt caching, sub-agents).
Lovable uses a fixed internal model — your selected GPT-5.4 mini will be ignored.
Replit Agent 3 uses its own model selection internally — your selected GPT-5.4 mini will be ignored.

Build a full stack around GPT-5.4 mini — Flowpicker shows compatibility warnings before you commit.