GPT-5.4 Pro
More exact GPT-5.4 tier for demanding professional reasoning and agent tasks
- Context
- 1.1M
- Released
- 7mo ago
- Input / 1M tokens
- $30.00
- Output / 1M tokens
- $180.00
Sources:models.devOpenRouter
Benchmarks
DeepSWE is the primary coding benchmark where an exact model evaluation exists. Other sourced evaluations appear below; missing scores are not zero.
Not evaluated in DeepSWE 1.1 as of 2026-09-22. Results are only shown for exact model versions in the official dataset. Official leaderboard.
No benchmark results have been matched to this model yet.
Specifications
- Developer
- OpenAI
- Model family
- GPT
- Type
- Language model
- Predecessor
- —
- Successor
- —
- Released
- March 5, 2026
- Model ID
openai/gpt-5.4-pro- Openness
- Closed
- License
- —
- Input modalities
- text, image, file
- Output modalities
- text
- Modality
- text+image+file->text
- Tokenizer
- GPT
- Max output tokens
- 128K
- Knowledge cutoff
- 2025-08-31
- Hugging Face
- —
- Source freshness
- Metadata synced Oct 8, 2026, 10:17 PM UTC
Additional source evidence
- Source model
- openai/gpt-5.4-pro
- Source family
- gpt-pro
- Last updated
- 2026-03-05
- Knowledge cutoff
- 2025-08-31
- Model context
- 1.1M
- Maximum input
- 922K
- Model output limit
- 128K
- Open weights
- No
- File attachments
- Yes
- Reasoning
- Yes
- Tool calling
- Yes
- Structured output
- No
- Temperature control
- No
Published benchmarks
| Benchmark | Score | Evaluation details | Evidence |
|---|---|---|---|
| GPQA Diamond | 94.4 · accuracy | 2026-04-23 | Source |
| Humanity's Last Exam | 42.7 · accuracy | no tools · 2026-04-23 | Source |
| Humanity's Last Exam | 58.7 · accuracy | with tools · 2026-04-23 | Source |
| BrowseComp | 89.3 · accuracy | 2026-04-23 | Source |
| GDPval | 82 · wins or ties | 2026-04-23 | Source |
| FrontierMath | 50 · accuracy | Tier 1-3 · 2026-04-23 | Source |
| FrontierMath | 38 · accuracy | Tier 4 · 2026-04-23 | Source |
| ARC-AGI-1 | 94.5 · accuracy | Verified · 2026-04-23 | Source |
| ARC-AGI-2 | 83.3 · accuracy | Verified · 2026-04-23 | Source |
| FinanceAgent | 61.5 · accuracy | 1.1 · 2026-04-23 | Source |
| GeneBench | 25.6 · accuracy | 2026-04-23 | Source |
Complete source record (JSON)
{
"id": "openai/gpt-5.4-pro",
"name": "GPT-5.4 Pro",
"description": "More exact GPT-5.4 tier for demanding professional reasoning and agent tasks",
"family": "gpt-pro",
"last_updated": "2026-03-05",
"knowledge": "2025-08-31",
"attachment": true,
"reasoning": true,
"tool_call": true,
"structured_output": false,
"temperature": false,
"modalities": {
"input": [
"text",
"image"
],
"output": [
"text"
]
},
"limit": {
"context": 1050000,
"input": 922000,
"output": 128000
},
"benchmarks": [
{
"name": "GPQA Diamond",
"score": 94.4,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"date": "2026-04-23"
},
{
"name": "Humanity's Last Exam",
"score": 42.7,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"variant": "no tools",
"date": "2026-04-23"
},
{
"name": "Humanity's Last Exam",
"score": 58.7,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"variant": "with tools",
"date": "2026-04-23"
},
{
"name": "BrowseComp",
"score": 89.3,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"date": "2026-04-23"
},
{
"name": "GDPval",
"score": 82,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "wins or ties",
"date": "2026-04-23"
},
{
"name": "FrontierMath",
"score": 50,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"dataset": "Tier 1-3",
"date": "2026-04-23"
},
{
"name": "FrontierMath",
"score": 38,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"dataset": "Tier 4",
"date": "2026-04-23"
},
{
"name": "ARC-AGI-1",
"score": 94.5,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"variant": "Verified",
"date": "2026-04-23"
},
{
"name": "ARC-AGI-2",
"score": 83.3,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"variant": "Verified",
"date": "2026-04-23"
},
{
"name": "FinanceAgent",
"score": 61.5,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"version": "1.1",
"date": "2026-04-23"
},
{
"name": "GeneBench",
"score": 25.6,
"source": "https://openai.com/index/introducing-gpt-5-5/",
"metric": "accuracy",
"date": "2026-04-23"
}
],
"release_date": "2026-03-05",
"open_weights": false
}Infrastructure
Provider availability
| Provider | Endpoint | Context | Input /M tokens | Output /M tokens | Cached input /M tokens | Cache write /M tokens | Uptime 1d | Data handling |
|---|---|---|---|---|---|---|---|---|
| openai/flex | 1.1M | $15.00 | $90.00 | — | — | — | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| openai/flexBatch pricing | 1.1M | $15.00 | $90.00 | — | — | — | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| azure | 1.1M | $30.00 | $180.00 | — | — | 100.00% | ZDR: YesTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| azureBatch pricing | 1.1M | $30.00 | $180.00 | — | — | 100.00% | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| openai | 1.1M | $30.00 | $180.00 | — | — | — | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 | |
| openaiBatch pricing | 1.1M | $30.00 | $180.00 | — | — | — | ZDR: NoTraining: Unknown Prompt logging: Unknown Retention: Unknown Checked 2026-10-15 |
Provider data:OpenRouter
Text provenance
EU · product-scopedOpenAI textGrain watermarking
In the EU, ChatGPT-generated text uses textGrain. OpenAI says API customers can opt in globally for select models, so watermarking depends on the product surface and configuration.
Official explanation