GPT-5.4 Mini
budgetOpenAI•Released on 2026-03-18
OpenAI's fastest small model, delivering 2x speed improvement over GPT-5 Mini while approaching flagship GPT-5.4 accuracy. Excels at coding, tool use, and multimodal tasks. Ideal for subagent architectures and high-volume workloads. 72.1% OSWorld accuracy (vs 75% GPT-5.4, 42% GPT-5 Mini).
79
Overall Score
Voice of the community
sample 50“GPT-5.4 mini approaches the accuracy of the flagship GPT-5.4 model on evaluations like SWE-Bench Pro, delivering one of the strongest performance-per-latency tradeoffs for developers.”
“Developers can handle routine coding tasks using GPT-5.4 mini for about one-third the cost, utilizing only 30 percent of the standard GPT-5.4 quota.”
Core Specs
400K
Context Window
16K
Max Output
ReasoningOpen Sourcetextimage
Scenario Scores
Pros & Cons
Sentiment85% +10% ·5% −
Pros
- +2x faster than GPT-5 Mini
- +Near-flagship accuracy at 1/3 cost
- +400K context window
- +Excellent for subagent tasks
- +Strong computer use capability
Cons
- −Less capable than full GPT-5.4
- −No reasoning tokens display
- −Not ideal for complex planning
Pricing
Input (per 1M tokens)$0.75
Output (per 1M tokens)$4.50
Free trial available
Updated on 2026-03-19
Get Started
1Visit the provider's website
2Create an account
3Start using the model
Benchmarks
sweProVerified68%
osWorldVerified72.1%
noteNear-flagship performance at 2x speed%