Compare Start here

OpenAILanguage modelSuperseded

GPT-5.4

Merged the Codex coding line back into the main GPT-5 model.

Compare head-to-head Race it in 3D

Overall Intelligence

#19

of 56
7.6 pts behind #1

0

races led

85

weeks at #1

81.1

score

Released: Mar 5, 2026; 8 mo ago
Context: 400K; 128K max output
Price: $2.5 / $15; per 1M in/out · $5.63 blended
Speed: 85 tok/s; 6.5s to first token
Modalities: text · image · audio · video; outputs text
Weights: Closed; API and apps

The 30-second version

Merged the Codex coding line back into the main GPT-5 model.

Key takeaways

1Merged the Codex coding line back into the main GPT-5 model.

Where it ranks

Current snapshot · illustrative scores · click any race to watch it

GPT-5.4Best on each axis

Overall

19Overall+7.6 pts

Coding

20Coding+10.1 pts 21Frontend & UI+62 Elo 19Backend & APIs+8.7 pts 20Distributed Systems+11.8 pts 20Full-Stack Apps+10.3 pts 22Debugging & Repair+11.5 pts 23Agentic Coding+12.6 pts 21Code Review & Security+9.7 pts 16Data Science & SQL+6.1 pts 23Mobile Apps+9.3 pts 21Terminal & DevOps+13.1 pts 17Competitive Programming+226

Hard Reasoning

17Hard Reasoning+8.0 pts 17Frontier Questions+15.1 pts 15Abstract & Novel Reasoning+10.2 pts 18Logic & Puzzles+7.3 pts 16Planning & Strategy+8.0 pts

Math

16Math+8.4 pts 15Competition Math+6.6 pts 15Research Math+10.1 pts 18Applied & Quantitative+7.5 pts

Science

17Science+8.7 pts 17Graduate Science QA+5.2 pts 19Biology & Medicine+8.4 pts 16Physics & Chemistry+6.6 pts 20Research Assistance+8.9 pts

Writing & Language

20Writing & Language+9.8 pts 18Creative Writing+64 Elo 19Technical Writing+8.5 pts 22Editing & Style+8.1 pts 23Multilingual+6.7 pts 15Instruction Following+6.1 pts 24Conversation & EQ+70 Elo

Multimodal

16Multimodal+13.6 pts 17Image Understanding+9.0 pts 20Documents & Charts+10.1 pts 19Video Understanding+17.5 pts 15Speech & Audio+10.4 pts

Tool Use & Agents

18Tool Use & Agents+10.1 pts 17Tool Use+7.4 pts 17Computer Use+14.8 pts 20Web Research+14.4 pts 17Long-Horizon+20.2 hrs

Long Context

23Long Context+12.3 pts 261M Recall+11.6 pts 22Long Reasoning+9.2 pts

Speed & Cost

37Speed & Cost+44.7 pts 26Output Speed+438 tok/s 43Latency+6.24s 30Cost-Efficiency+46.4 pts

Trust & Factuality

21Trust & Factuality+11.2 pts 24Factuality+9.9 pts 19Honesty & Calibration+8.6 pts

Overall Intelligence over time

#1 frontierGPT-5.4Spark OpenKimi K3

Update history

The distilled version first; open any release for the deeper breakdown.

The GPT-5 family

GPT-5August 2025 GPT-5.1November 2025 GPT-5.2December 2025 GPT-5.4March 2026 GPT-5.5April 2026

60-second briefing

GPT-5.4, explained

AI video summary · placeholder render from the auto-written script

From OpenAI's own guide, distilled

Prompting guides for GPT-5 and later models

Reasoning models do their own step-by-step thinking; give goals and constraints rather than scripting the chain of thought.
Use the reasoning-effort and verbosity controls instead of prompt tricks to trade depth for speed.
Contradictory instructions waste reasoning tokens; audit system prompts for conflicts.
For agents, define stop conditions and how eager the model should be to keep going without asking.
Structured outputs (JSON schema) beat asking politely for a format.

All official guides, distilled →

About OpenAI

Maker of ChatGPT and the GPT series, pursuing general-purpose AI at consumer scale.

Optimising for

Frontier reasoning and math, broad consumer usefulness inside ChatGPT, and tiered models (Astra, Sol, Luna) that trade depth for price and speed.

San Francisco, US · founded 2015

Closest rivals

GPT-5.4 is 19th in overall intelligence; leader is Claude Opus 5.5.