GPT-5.4
Parameters
Undisclosed
Context Length
272K
Modality
Multimodal
Architecture
Undisclosed
License
Proprietary
Release Date
5 Mar 2026
Knowledge Cutoff
-
API Pricing (per 1M)
Input: $2.50 · Output: $15.00
Evaluation Benchmarks
Rank
#22
| Benchmark | Score | Rank |
|---|---|---|
Data Analysis | max 0.79 | 8 |
Mathematics | max 0.94 | 11 |
Graduate-Level QA | 0.928 | 11 |
LiveBench Average | max 0.78 | 12 |
General | max 0.78 | 13 |
Reasoning | max 0.88 | 16 |
General Text | high 1476 Standard 1466 | 23 36 |
Agentic Coding | max 0.54 | 23 |
Agentic Index | max 0.44 | 25 |
Coding | max 0.78 | 26 |
Agent Arena | high 0.02 | 27 |
Intelligence Index | max 0.39 low 0.28 Standard 0.18 | 37 83 129 |
Coding Index | max 0.71 | 41 |
Web Development | high 1463 medium 1443 Standard 1387 | 44 49 64 |
Rankings
Overall Rank
#22
Coding Rank
#38
About GPT-5.4
GPT-5.4 is OpenAI's most capable and efficient frontier model for professional work. It brings together advances in reasoning, coding, and agentic workflows into a single model. Features industry-leading coding capabilities from GPT-5.3-Codex, native state-of-the-art computer-use capabilities, and improved tool use across large ecosystems. Excels at professional tasks involving spreadsheets, presentations, and documents. Achieves 83.0% on GDPval, 75.0% on OSWorld-Verified, 82.7% on BrowseComp, 57.7% on SWE-Bench Pro, and 81.2% on MMMU Pro. Supports up to 272K context (1M experimental) and delivers most token-efficient reasoning yet.
Technical Specifications
Architecture specifications are undisclosed for proprietary models.
Attention
Attention Structure
Multi-Head Attention
Attention Heads
Undisclosed
Key-Value Heads
Undisclosed
Attention Head Dimension
Undisclosed
Position Embedding
Absolute Position Embedding
RoPE Theta
Undisclosed
Sliding Window Attention
Undisclosed
Sliding Window Size
Undisclosed
Sliding Window Ratio
Undisclosed
Linear Attention
Undisclosed
Linear Attention Ratio
Undisclosed
Normalization
Undisclosed
Activation Function
Undisclosed
Dimensions
Hidden Dimension Size
Undisclosed
Number of Layers
Undisclosed
FFN Intermediate Size (Dense)
Undisclosed
Multi-Token Prediction Heads
Undisclosed
Tokenizer
Vocabulary Size
Undisclosed
Resources
About GPT-5.4
GPT-5.4 is OpenAI's most capable and efficient frontier model for professional work, combining the industry-leading coding capabilities of GPT-5.3-Codex with major advances in reasoning, computer use, and agentic workflows. It introduces native computer-use capabilities, tool search for large tool ecosystems, substantially improved knowledge work (spreadsheets, presentations, documents), and is OpenAI's most factual and token-efficient reasoning model. Supports up to 1M context tokens in Codex. Released March 5, 2026.