GPT-5.4
GPTBalanced
ThinkingTool UseVisionStructured Output
About this model
GPT-5.4 is OpenAI’s latest frontier model, unifying the Codex and GPT lines into a single system. It features a 1M+ token context window (922K input, 128K output) with support for...
Performance Tier
Balanced
GPT-5.4 is a balanced model from GPT : strong performance at a reasonable price.
Strong cost-performance ratio. Reliable for most professional use cases without premium pricing.
Pricing
This model is included in Elosia plansHigh
High pricing. Reserve this level for tasks that demand maximum quality.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $2.50 |
| Output (completion) | $15.00 |
| Cache read | $0.250 |
Capabilities
Context Length1.1M
Max Output Tokens128K
TokenizerGPT
Inputtext, image, file
Outputtext
Release DateMarch 5, 2026
Benchmarks
General Intelligence
MMLU
Not reported
Mathematics
MATH-500
Not reported
Programming
HumanEval
Not reported
Reasoning
ARC-AGI-2
73.3%
Humanity's Last Exam
39.8%
Recommended Use Cases
CodingAnalysisResearchGeneral ChatData Extraction
Strengths
- Unifies the Codex and GPT lines into a single frontier model
- 1M-token context window for large-scale document analysis
- 33% fewer hallucinations compared to GPT-5.2
- Native computer use — clicks, types and navigates software via screenshots
Limitations
- Premium pricing tier
- Surpassed by GPT-5.5 on agentic benchmarks (Terminal-Bench 2.0, ARC-AGI-2) and long-context retrieval
- Reasoning overhead makes it slower than GPT-5.4 Mini/Nano for latency-sensitive tasks
Resources
This model may use your data for training