GPT-5.4
GPT-5.4 is an API-only (closed weights) language model from OpenAI, released March 5, 2026. The context window is 1.1M tokens. On the Epoch Capabilities Index it scores 157.8, rank 27 of 256 rated models. API price: $2.5 input / $15 output per million tokens.
Check it against your own PC →
Lab
OpenAI
Released
March 5, 2026
Weights
API-only (closed weights)
Parameters
–
Context
1.1M
Price / M tokens
$2.5 / $15 in / out
Epoch ECI
157.8 #27
Strengths by category
ModelTape's 0–100 fit on Epoch AI benchmark results.
| Category | Score | Based on |
|---|---|---|
| Reasoning | 26.8 | Chess Puzzles, Mystery Game Puzzles, Humanity's Last Exam |
| Math | 61.2 | ProofBench, OTIS Mock AIME 2024-2025, FrontierMath-Tiers-1-3-v2-Private, FrontierMath-Tier-4-v2-Private |
| Coding | 55.2 | GSO-Bench, MirrorCode, Terminal Bench, SWE-bench Verified, DeepSWE |
| Knowledge | 75.7 | GPQA diamond, SimpleQA Verified |
Benchmark results
| Benchmark (Epoch AI) | Score |
|---|---|
| OTIS Mock AIME | 97.8% |
| GPQA Diamond | 94.9% |
| FrontierMath (Tiers 1–3) | 82.7% |
| Terminal-Bench | 81.8% |
| SWE-bench Verified | 76.9% |
| ProofBench | 56.0% |
| DeepSWE | 51.8% |
| FrontierMath-Tier-4-v2-Private | 49.0% |
| Chess Puzzles | 45.3% |
| SimpleQA Verified | 45.1% |
| Humanity's Last Exam | 36.2% |
| GSO-Bench | 31.4% |
| Mystery Game Puzzles | 30.6% |
| EBR-bench | 25.4% |
| MirrorCode | 23.3% |
| Furniture Assembly | 10.7% |
Links
Announcement · Epoch AI benchmarks
Models with a similar score
Muse Wonder · Kimi K3.1 Pro · GPT-5.6 Terra · Claude Haiku 5.5 · GPT-5.4 Pro · GPT-5.5 · Claude Opus 4.8 · DeepSeek V4.1 Flash 1017
More from OpenAI
GPT-6.1 Terra · GPT-6.1 Sol · GPT-6 Sol · GPT-6 Luna · GPT-6 Astra · GPT-5.6 Cyber · Daybreak Red · Daybreak Blue