OpenAI model product
GPT-5.4
4 is an OpenAI large language model with coding, native computer-use, vision, tool search, and reasoning capabilities.
Updated Sep 8, 2026. Default version: GPT-5.4
LLMBoard Score75.2GPT-5.4
Coverage40%23 benchmark families
Context window1MTokens
Official input price$2.5OpenAI API
GPT-5.4 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
GPT-5.4 Benchmark Results
Benchmark scores for GPT-5.4.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkGraphwalks parents <128k | Score89.80% | Rank01 | Participants11 | Percentile100.00% | EvidenceC | Evaluated |
| BenchmarkGraphwalks BFS <128k | Score93.00% | Rank02 | Participants11 | Percentile90.00% | EvidenceC | Evaluated |
| BenchmarkARC-AGI | Score93.70% | Rank03 | Participants9 | Percentile75.00% | EvidenceC | Evaluated |
| BenchmarkLiveBench | Score80.28% | Rank03 | Participants38 | Percentile94.59% | EvidenceB | Evaluated |
| BenchmarkTau2 Telecom | Score98.90% | Rank03 | Participants36 | Percentile94.29% | EvidenceC | Evaluated |
| BenchmarkARC-AGI v2 | Score73.30% | Rank04 | Participants18 | Percentile82.35% | EvidenceC | Evaluated |
| BenchmarkFrontierMath | Score47.60% | Rank04 | Participants17 | Percentile81.25% | EvidenceC | Evaluated |
| BenchmarkGraphwalks parents >128k | Score32.40% | Rank04 | Participants7 | Percentile50.00% | EvidenceC | Evaluated |
| BenchmarkFinance Agent | Score56.00% | Rank06 | Participants8 | Percentile28.57% | EvidenceC | Evaluated |
| BenchmarkTerminal-Bench 2.0 | Score75.10% | Rank06 | Participants51 | Percentile90.00% | EvidenceC | Evaluated |
| BenchmarkGraphwalks BFS >128k | Score21.40% | Rank08 | Participants11 | Percentile30.00% | EvidenceC | Evaluated |
| BenchmarkOmniDocBench 1.5 | Score89.10% | Rank08 | Participants18 | Percentile58.82% | EvidenceC | Evaluated |
| BenchmarkLM Arena Search Factuality | Score1,211.91 rating | Rank08 | Participants28 | Percentile74.07% | EvidenceA | Evaluated |
| BenchmarkLegal Agent Benchmark | Score0.40% | Rank09 | Participants13 | Percentile33.33% | EvidenceB | Evaluated |
| BenchmarkMMMU-Pro | Score81.20% | Rank09 | Participants69 | Percentile88.24% | EvidenceC | Evaluated |
| BenchmarkLM Arena Text Factuality | Score1,485.06 rating | Rank09 | Participants121 | Percentile93.33% | EvidenceA | Evaluated |
| BenchmarkFrontierSWE | Score54.00% | Rank10 | Participants16 | Percentile40.00% | EvidenceB | Evaluated |
| BenchmarkLM Arena Vision | Score1,298.31 rating | Rank11 | Participants107 | Percentile90.57% | EvidenceA | Evaluated |
| BenchmarkGPQA | Score92.80% | Rank12 | Participants247 | Percentile95.53% | EvidenceC | Evaluated |
| BenchmarkOSWorld-Verified | Score75.00% | Rank12 | Participants24 | Percentile52.17% | EvidenceC | Evaluated |
| BenchmarkLM Arena Document | Score1,470.14 rating | Rank12 | Participants33 | Percentile65.63% | EvidenceA | Evaluated |
| BenchmarkLM Arena Document Style Control | Score1,469.34 rating | Rank12 | Participants33 | Percentile65.63% | EvidenceA | Evaluated |
| BenchmarkLM Arena Search Style Control | Score1,194.52 rating | Rank13 | Participants28 | Percentile55.56% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Steerability | Score4.16% | Rank14 | Participants49 | Percentile72.92% | EvidenceA | Evaluated |
| BenchmarkToolathlon | Score54.60% | Rank15 | Participants41 | Percentile65.00% | EvidenceC | Evaluated |
| BenchmarkLM Arena Agent Bash Recovery Steps | Score8.19% | Rank15 | Participants49 | Percentile70.83% | EvidenceA | Evaluated |
| BenchmarkLM Arena Agent Tool Hallucination | Score0.78% | Rank15 | Participants49 | Percentile70.83% | EvidenceA | Evaluated |
| BenchmarkLM Arena Search | Score1,197.16 rating | Rank15 | Participants28 | Percentile48.15% | EvidenceA | Evaluated |
| BenchmarkLM Arena Vision Style Control | Score1,280.73 rating | Rank15 | Participants107 | Percentile86.79% | EvidenceA | Evaluated |
| BenchmarkBrowseComp | Score82.70% | Rank20 | Participants63 | Percentile69.35% | EvidenceC | Evaluated |
GPT-5.4 Arena Results
Preference and agent-evaluation results for the default version.
Sort by
Arena | Category | Rank | Rating / score | Votes | Observations | Result date |
|---|
| Arenatext factuality | Categoryfrench | Rank03 | Rating / score1,523.89 | Votes1,808 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry business and management and financial operations | Rank03 | Rating / score1,501.45 | Votes10,593 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry mathematical | Rank03 | Rating / score1,498.33 | Votes2,638 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryjapanese | Rank03 | Rating / score1,483.24 | Votes437 | ObservationsN/A | Result date |
| Arenatext factuality | Categorygerman | Rank04 | Rating / score1,462.68 | Votes613 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry legal and government | Rank04 | Rating / score1,513.49 | Votes3,950 | ObservationsN/A | Result date |
| Arenatext factuality | Categorymulti turn | Rank04 | Rating / score1,501.84 | Votes9,515 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryrussian | Rank04 | Rating / score1,499.40 | Votes5,201 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryexpert | Rank06 | Rating / score1,528.75 | Votes4,756 | ObservationsN/A | Result date |
| Arenatext factuality | Categorypolish | Rank06 | Rating / score1,473.44 | Votes854 | ObservationsN/A | Result date |
| Arenatext factuality | Categorychinese | Rank07 | Rating / score1,533.24 | Votes2,860 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryhard prompts | Rank07 | Rating / score1,510.70 | Votes39,196 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry software and it services | Rank07 | Rating / score1,519.51 | Votes23,061 | ObservationsN/A | Result date |
| Arenatext factuality | Categorykorean | Rank07 | Rating / score1,437.48 | Votes665 | ObservationsN/A | Result date |
| Arenatext factuality | Categorymath | Rank07 | Rating / score1,492.65 | Votes2,612 | ObservationsN/A | Result date |
| Arenatext factuality | Categorynon english | Rank07 | Rating / score1,475.46 | Votes33,012 | ObservationsN/A | Result date |
| Arenavision | Categorydiagram | Rank07 | Rating / score1,322.92 | Votes6,270 | ObservationsN/A | Result date |
| Arenavision | Categoryhomework | Rank07 | Rating / score1,335.39 | Votes3,053 | ObservationsN/A | Result date |
| Arenasearch factuality | Categoryoverall | Rank08 | Rating / score1,211.91 | Votes110,062 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryhard prompts english | Rank08 | Rating / score1,513.10 | Votes17,259 | ObservationsN/A | Result date |
| Arenatext | Categoryjapanese | Rank09 | Rating / score1,486.98 | Votes644 | ObservationsN/A | Result date |
| Arenatext factuality | Categorycoding | Rank09 | Rating / score1,533.26 | Votes15,192 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryenglish | Rank09 | Rating / score1,488.81 | Votes27,347 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryindustry medicine and healthcare | Rank09 | Rating / score1,503.73 | Votes3,573 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryinstruction following | Rank09 | Rating / score1,484.83 | Votes18,978 | ObservationsN/A | Result date |
| Arenatext factuality | Categoryoverall | Rank09 | Rating / score1,485.06 | Votes60,522 | ObservationsN/A | Result date |
| Arenavision | Categorychinese | Rank09 | Rating / score1,357.58 | Votes1,282 | ObservationsN/A | Result date |
| Arenavision | Categoryenglish | Rank09 | Rating / score1,297.18 | Votes9,529 | ObservationsN/A | Result date |
| Arenavision style control | Categoryhomework | Rank09 | Rating / score1,329.48 | Votes3,053 | ObservationsN/A | Result date |
| Arenatext | Categoryfrench | Rank10 | Rating / score1,494.84 | Votes2,285 | ObservationsN/A | Result date |
GPT-5.4 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $2.5 input, $15 output per 1M
- Official provider
- OpenAI
- Lowest third-party
- From $0.75 input, $6 output per 1M via Xpersona
- Tracked offerings
- 44
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderUnoRouter | Provider model IDgpt-5.4:free | Regionglobal | Input / 1MN/A | Output / 1MN/A | Context1.1M | Updated |
| ProviderXpersona | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$0.75 | Output / 1M$6 | Context1.1M | Updated |
| ProviderUnoRouter | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$1.8 | Output / 1M$10.8 | Context1.1M | Updated |
| ProviderPoe | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.2 | Output / 1M$14 | Context1.1M | Updated |
| ProviderAI-ROUTER | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderOrcaRouter | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderNanoGPT | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderImpossibl | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderOpenRouter | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderDaoXE | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderCrossModel | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderDatabricks | Provider model IDdatabricks-gpt-5-4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderOpenAI | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderOpper | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderLLM Gateway | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderVivgrid | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context400K | Updated |
| ProviderNEAR AI Cloud | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderMerge Gateway | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderCloudflare AI Gateway | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1M | Updated |
| ProviderPerplexity Agent | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderSAP AI Core | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderFreeModel | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderOfox | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderNeon | Provider model IDgpt-5-4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderAzure Cognitive Services | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| Provider302.AI | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderOpenCode Zen | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderAzure | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderAIHubMix | Provider model IDgpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
| ProviderVercel AI Gateway | Provider model IDopenai/gpt-5.4 | Regionglobal | Input / 1M$2.5 | Output / 1M$15 | Context1.1M | Updated |
GPT-5.4 Runtime Performance
Provider-specific output speed and catalog latency for GPT-5.4. Runtime does not affect the capability score.
Sort by
Provider | Output Speed | Catalog Latency | Max Input | Max Output | Updated |
|---|
| ProviderOpenAI | Output Speed50.00 tok/s | Catalog Latency3.00 s | Max Input1M | Max Output128K | Updated |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
GPT-5.4 Specifications
Technical details for the model's default version.
- Version
- GPT-5.4
- Released
- Mar 5, 2026
- Knowledge cutoff
- Unknown
- Parameters
- N/A
- Context window
- 1M
- Max output
- 128K
- Inputs
- image, text
- Outputs
- text
- Open weights
- No
- License
- Proprietary
GPT-5.4 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionGPT-5.4 | Released | LLMBoard75.18 | ParametersN/A | Context1M | Max output128K | Open weightsNo | LicenseProprietary |
Models similar to GPT-5.4
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
OP
GPT-5.5-Pro
OpenAI
OP
GPT-5.6-Luna
OpenAI
OP
GPT-5.2-Pro
OpenAI
OP
GPT-5.2
OpenAI
OP
GPT-5.5
OpenAI
OP
GPT-5.6-Terra
OpenAI
What is GPT-5.4?
Key information about GPT-5.4 and its available data.
4 is an OpenAI large language model with up to 1M tokens of context and full-resolution vision processing. It supports coding, native computer-use, tool search for large tool ecosystems, and reasoning across spreadsheets, presentations, and documents.
Data as of 2026-09-08.
FAQ
Common questions about GPT-5.4.
When was GPT-5.4 released?
GPT-5.4's default version was released on Mar 5, 2026.
How much does GPT-5.4 cost?
GPT-5.4's official API price is $2.5 per million input tokens and $15 per million output tokens via OpenAI. The lowest tracked third-party offer starts at $0.75 input and $6 output via Xpersona.
Who created GPT-5.4?
GPT-5.4 was created by OpenAI.
What is the context window for GPT-5.4?
The default version has a 1M token context window.
Is GPT-5.4 open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer GPT-5.4?
44 provider offerings are linked to the default version.