GPT-5.4 Models by OpenAI
OpenAIProprietary
6 models2026Up to 1.05m ctxFrom $0.2/1M input
Details
ResearcherOpenAI
LicenseProprietary
Commercial useCommercial use: conditional
Models6
Released2026
Max context1.05m
Capabilities
Vision5 of 6 models
Multimodal5 of 6 models
Reasoning4 of 6 models
Function Calling5 of 6 models
Tool Use5 of 6 models
Structured Outputs5 of 6 models
Code Execution5 of 6 models
Links
WebsiteAbout
GPT-5.4 is OpenAI's frontier reasoning model family, designed for complex real-world tasks including spreadsheet editing, frontend code generation, document understanding, image analysis, tool use, and deep web research. Released March 5, 2026, as the successor to GPT-5.2 and GPT-5.3-Codex. The family includes full, mini, nano, and cybersecurity-specialized variants.
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
5 in view1 retired
GPT-5.4-CyberCurrent
Use when the workload needs cybersecurity, reasoning, and multimodal inputs.
2026-04cybersecurityreasoningmultimodal inputs
GPT-5.4Current
Use when the workload needs 1.05m context, reasoning, and tool use.
2026-031.05m contextreasoningtool use
GPT-5.4 MiniCurrent
Use when the workload needs 400k context, reasoning, and tool use.
2026-03400k contextreasoningtool use
GPT-5.4 NanoCurrent
Use when the workload needs 400k context, tool use, and function calling.
2026-03400k contexttool usefunction calling
GPT-5.4 ProCurrent
Use when the workload needs 1.05m context, reasoning, and tool use.
2026-031.05m contextreasoningtool use
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| GPT-5.4-Cyber | Use when the workload needs cybersecurity, reasoning, and multimodal inputs. | 2026-04 | cybersecurityreasoningmultimodal inputs | Current |
| GPT-5.4 | Use when the workload needs 1.05m context, reasoning, and tool use. | 2026-03 | 1.05m contextreasoningtool use | Current |
| GPT-5.4 Mini | Use when the workload needs 400k context, reasoning, and tool use. | 2026-03 | 400k contextreasoningtool use | Current |
| GPT-5.4 Nano | Use when the workload needs 400k context, tool use, and function calling. | 2026-03 | 400k contexttool usefunction calling | Current |
| GPT-5.4 Pro | Use when the workload needs 1.05m context, reasoning, and tool use. | 2026-03 | 1.05m contextreasoningtool use | Current |
Release Timeline
3 release groups2026-04
1 current
GPT-5.4-Cyber
Currentcybersecurityreasoningmultimodal inputs
2026-03
4 current
GPT-5.4
Current1.05m contextreasoningtool use
GPT-5.4 Mini
Current400k contextreasoningtool use
GPT-5.4 Nano
Current400k contexttool usefunction calling
GPT-5.4 Pro
Current1.05m contextreasoningtool use
2026-02
1 retired
GPT-5.3-Codex-Spark
Replacedcode131k contexttool use
Replaced By
Replaced
Keep for legacy integrations; evaluate GPT-5.3-Codex before new work.
Specifications(6 models)
| Model | Released | Context | Vision | Multimodal | Reasoning | Fn Calling | Tool Use | Structured Outputs | Code Exec |
|---|---|---|---|---|---|---|---|---|---|
| GPT-5.4-Cyber | 2026-04 | — | Yes | Yes | Yes | No | No | No | No |
| GPT-5.4 | 2026-03 | 1.05m | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| GPT-5.4 Mini | 2026-03 | 400k | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| GPT-5.4 Nano | 2026-03 | 400k | Yes | Yes | No | Yes | Yes | Yes | Yes |
| GPT-5.4 Pro | 2026-03 | 1.05m | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
Available From(4 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| GPT-5.4 Nano | OpenAI API | $0.2 | $1.25 | Serverless |
| GPT-5.4 Nano | OpenRouter | $0.2 | $1.25 | Serverless |
| GPT-5.4 Nano | Vercel AI Gateway | $0.2 | $1.25 | Serverless |
| GPT-5.4 Mini | OpenAI API | $0.75 | $4.5 | Serverless |
| GPT-5.4 Mini | OpenRouter | $0.75 | $4.5 | Serverless |
| GPT-5.4 Mini | Vercel AI Gateway | $0.75 | $4.5 | Serverless |
| GPT-5.4 | OpenAI API | $2.5 | $15 | Serverless |
| GPT-5.4 | OpenRouter | $2.5 | $15 | Serverless |
| GPT-5.4 | Vercel AI Gateway | $2.5 | $15 | Serverless |
| GPT-5.4 | AWS Bedrock | $2.75 | $16.5 | Serverless |
| GPT-5.4 Pro | OpenRouter | $30 | $180 | Serverless |
| GPT-5.4 Pro | OpenAI API | $30 | $180 | Serverless |
| GPT-5.4 Pro | Vercel AI Gateway | $30 | $180 | Serverless |
Popular comparisons in this family
Comparisons
- Qwen3.6-Max vs GPT-5.4
- Claude Sonnet 5 vs GPT-5.4
- Claude Opus 4.8 vs GPT-5.4
- Claude Opus 4.7 vs GPT-5.4
- Composer 2.5 vs GPT-5.4
- GPT-5.5 vs GPT-5.4
- GPT-5.5 Pro vs GPT-5.4
Frequently Asked Questions
- What is GPT-5.4 used for?
- GPT-5.4 is used for cybersecurity, code, and vision and multimodal work. The family description and listed model capabilities point to those workloads as the best fit.
- How does GPT-5.4 compare to GPT Realtime 2?
- GPT-5.4 by OpenAI is strongest where you need cybersecurity, while GPT Realtime 2 by OpenAI is the closest related family to check for realtime voice. GPT-5.4 has 6 listed variants and reaches up to 1.05m context, while GPT Realtime 2 reaches up to 131k context, so compare the specs and pricing tables before choosing a production model.
- Which GPT-5.4 model should I use?
- For the lowest listed input price, start with GPT-5.4 Nano through OpenAI API at $0.2/1M input tokens. For the most capable/latest local choice, evaluate GPT-5.4 with 1.05m context and reasoning, tool use, function calling, structured outputs, and multimodal inputs.
