Data as of October 17, 2026 · How the score is built
GPT-5.4
Decision readingGPT-5.4 scores 66.7 out of 100 and ranks #30 of 224. This profile shows 42 source-displayable benchmark rows; its strongest eligible category is Knowledge at #20. API pricing is $2.5 input and $15 output per million tokens, with cached input at $0.25.
Released Mar 5, 2026 — see all recent releases
Decision snapshot
Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.
Capability
66.7/100
field median 49.6#30 of 224 ranked models
Public
#30of 224
Verified #21 of 88
Price
$2.50input / $15 output
input median $0.75cached $0.25 · blended $8.75
Speed
89tok/s
field median 92 tok/sFirst token 157.51 s
Context
1.05Mtokens
field median 256,000Maximum output length is tracked separately
Strongest published evidence
Knowledge ranks #20. Particularly effective for knowledge-intensive tasks like research, analysis, and factual Q&A.
Validate before choosing
42 published rows leave some tracked benchmark slots empty. Coding is its lowest eligible category at #48.
Source-linked · 42 displayable benchmark rows
Category score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #44 of 128Percentile 66thWeight 22%15 benchmarksMixed sources | 52.1 | 15 benchmarks | Mixed sources | |||
| CodingRank #48 of 150Percentile 68thWeight 20%6 benchmarksMixed sources | 47.7 | 6 benchmarks | Mixed sources | |||
| ReasoningRank #23 of 28Percentile 19thWeight 17%3 benchmarksVerified | 64.2 | 3 benchmarks | Verified | |||
| MultimodalRank #30 of 54Percentile 45thWeight 12%9 benchmarksMixed sources | 73.1 | 9 benchmarks | Mixed sources | |||
| KnowledgeRank #20 of 184Percentile 90thWeight 12%7 benchmarksMixed sources | 69.1 | 7 benchmarks | Mixed sources | |||
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| Inst. FollowingWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MathRank Not rankedWeight 5%2 benchmarksVerified | 82.0 | 2 benchmarks | Verified |
42 of 681 tracked benchmark slots have displayable evidence · bars run 0–100
Coverage detailsHow much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic14/15 verified
- Coding5/6 verified
- Reasoning3/3 verified
- Multimodal3/9 verified
- Knowledge3/7 verified
- MultilingualNot measured
- Inst. FollowingNot measured
- Math2/2 verified
Capability shape
Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.
GPT-5.4 category percentile values
- Agentic66th percentile
- Coding68th percentile
- Reasoning19th percentile
- Multimodal45th percentile
- Knowledge90th percentile
- MultilingualNot eligible
- Instruction followingNot eligible
- MathNot eligible
The dashed outline is median of 6 nearest peers.
Eligible category ranks
- Agentic#44/128
- Coding#48/150
- Reasoning#23/28
- Multimodal#30/54
- Knowledge#20/184
- MultilingualNot ranked
- Inst. FollowingNot ranked
- MathNot ranked
Benchmark ledger
Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.
Coding6 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| SWE-bench Pro | Score57.7% | Versus best verified row Best verified: Claude Opus 5.5 · 89.9% | Gap32.2 behind | Weight26% ref. weight | Provider exact |
| Terminal-Bench 2.0 | Score75.1% | Versus best verified row Best verified: GPT-5.5 · 82.0% | Gap6.9 behind | WeightScored in Agentic | Provider exact |
| LiveCodeBench Pro | Score87.5% | Versus best verified row Best verified: Sakana Fugu-Ultra · 90.8% | Gap3.3 behind | WeightDisplay only | Secondary exact |
| React Native Evals | Score85.3% | Versus best verified row Best verified: Composer 2 · 96.1% | Gap10.8 behind | WeightDisplay only | Benchmark exact |
| Vibe Code BenchVibe Code Bench v1.1 | Score67.42% | Versus best verified row Best verified: Gemini 4 Argon · 91.90% | Gap24.5 behind | WeightDisplay only | Benchmark exact |
| PostTrainBench v1.1 | Score19.0% | Versus best verified row Best verified: Claude Opus 5.5 · 49.3% | Gap30.3 behind | WeightDisplay only | Benchmark exact |
Agentic15 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| BrowseComp | Score82.7% | Versus best verified row Best verified: Atria Dawn Preview · 92.5% | Gap9.8 behind | Weight8% ref. weight | Provider exact |
| Terminal-Bench 2.0 | Score75.1% | Versus best verified row Best verified: GPT-5.5 · 82% | Gap6.9 behind | Weight7% ref. weight | Provider exact |
| OSWorld-Verified | Score75% | Versus best verified row Best verified: Qwen3.8 Max · 86.1% | Gap11.1 behind | Weight7% ref. weight | Provider exact |
| JobBench | Score38.9% | Versus best verified row Best verified: Muse Wonder Full · 75.3% | Gap36.4 behind | Weight6% ref. weight | Benchmark exact |
| MCP Atlas | Score70.6% | Versus best verified row Best verified: Muse Spark Open · 89.2% | Gap18.6 behind | Weight4% ref. weight | Provider exact |
| Toolathlon | Score54.6% | Versus best verified row Best verified: Muse Spark 1.1 · 75.6% | Gap21 behind | Weight3% ref. weight | Provider exact |
| DeepSearchQA | Score73.6% | Versus best verified row Best verified: Atria Dawn Preview · 96.0% | Gap22.4 behind | Weight2% ref. weight | Secondary exact |
| CyberGym | Score79.0% | Versus best verified row Best verified: MiMo-V2.6-Flash · 95.1% | Gap16.1 behind | WeightDisplay only | Provider exact |
| τ²-bench resultsτ²-Bench Tool-Agent-User Evaluation | Score98.9% | Versus best verified row Best verified: GPT-5.4 · 98.9% | GapBest verified | WeightDisplay only | Provider exact |
| Claw-Eval | Score60.3% | Versus best verified row Best verified: Ornith-1.5-397B · 81.4% | Gap21.1 behind | WeightDisplay only | Benchmark exact |
| Gert LabsGert Labs Composite Game Benchmark | Score64.89% | Versus best verified row Best verified: Claude Opus 4.8 · 72.97% | Gap8.1 behind | WeightDisplay only | Benchmark exact |
| ResearchClawBench | Score15.3% | Versus best verified row Best verified: Claude Opus 4.8 · 21.1% | Gap5.8 behind | WeightDisplay only | Benchmark exact |
| ExploitGym | Score6.0% | Versus best verified row Best verified: GPT-6 Astra · 42.4% | Gap36.4 behind | WeightDisplay only | Benchmark exact |
| ApprenticeBenchApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable job | Score11% | Versus best verified row Best verified: Claude Fable 5.1 · 72% | Gap61 behind | WeightDisplay only | Benchmark exact |
| MazeBenchMazeBench: visual spatial reasoning in a 3D open world | Score0% | Versus best verified row Best verified: GPT-6 Astra · 23% | Gap23 behind | WeightDisplay only | Benchmark exact |
Reasoning3 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| ARC-AGI-2Abstraction and Reasoning Corpus for AGI v2 | Score74.0% | Versus best verified row Best verified: GPT-6 Astra · 95% | Gap21 behind | WeightWeighted 25% | Benchmark exact |
| ARC-AGI-3Abstraction and Reasoning Corpus for AGI v3 | Score0.2% | Versus best verified row Best verified: GPT-6 Astra · 62.7% | Gap62.5 behind | WeightWeighted 15% | Benchmark exact |
| ARC-AGI-1ARC-AGI-1 Semi-Private Evaluation | Score93.67% | Versus best verified row Best verified: GPT-6 Astra · 98.50% | Gap4.8 behind | WeightDisplay only | Benchmark exact |
Multimodal9 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| MMMU-ProMassive Multi-discipline Multimodal Understanding Pro | Score81.2% | Versus best verified row Best verified: Gemini 3.5 Flash · 83.6% | Gap2.4 behind | WeightWeighted 40% | Provider exact |
| OfficeQA Pro | Score53.2% | Versus best verified row Best verified: Claude Opus 5.5 · 67.7% | Gap14.5 behind | WeightWeighted 25% | Provider exact |
| CharXivCharXiv Reasoning | Score82.8% | Versus best verified row Best verified: Qwen3.8 Max · 93.5% | Gap10.7 behind | WeightWeighted 20% | Secondary exact |
| MMMU-Pro w/ PythonMMMU-Pro with Python | Score82.1% | Versus best verified row Best verified: GPT-5.6 Sol · 84.6% | Gap2.5 behind | WeightDisplay only | Provider exact |
| ERQA | Score65.4% | Versus best verified row Best verified: Qwen3.8 Max · 77.8% | Gap12.4 behind | WeightDisplay only | Secondary exact |
| SimpleVQA | Score61.1% | Versus best verified row Best verified: Qwen3.7 Plus · 81.7% | Gap20.6 behind | WeightDisplay only | Secondary exact |
| ScreenSpot Pro | Score85.4% | Versus best verified row Best verified: GPT-6 Astra · 92.7% | Gap7.3 behind | WeightDisplay only | Secondary exact |
| ZeroBench | Score41.0% | Versus best verified row Best verified: Muse Spark Open · 50.0% | Gap9 behind | WeightDisplay only | Secondary exact |
| MedXpertQA (MM)MedXpertQA Multimodal | Score77.1% | Versus best verified row Best verified: Qwen3.8 Max · 80.4% | Gap3.3 behind | WeightDisplay only | Secondary exact |
Knowledge7 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| HLEHumanity's Last Exam | Score52.1% | Versus best verified row Best verified: Claude Fable 5.1 · 65% | Gap12.9 behind | Weight44% ref. weight | Provider exact |
| HLE w/o toolsHumanity's Last Exam without tools | Score39.8% | Versus best verified row Best verified: Claude Opus 5.5 · 64.4% | Gap24.6 behind | Weight7% ref. weight | Provider exact |
| GPQAGraduate-Level Google-Proof Q&A | Score92.8% | Versus best verified row Best verified: GPT-6 Astra · 96% | Gap3.2 behind | Weight3% ref. weight | Reported |
| GPQA-DGPQA Diamond | Score92.8% | Versus best verified row Best verified: GPT-6 Astra · 96.0% | Gap3.2 behind | WeightDisplay only | Secondary exact |
| HealthBench Hard | Score40.1% | Versus best verified row Best verified: Muse Spark · 42.8% | Gap2.7 behind | WeightDisplay only | Secondary exact |
| MedXpertQA (Text)MedXpertQA Text | Score59.6% | Versus best verified row Best verified: Muse Spark · 52.6% | Gap7 behind | WeightDisplay only | Secondary exact |
| HealthBench Professional | Score48.1% | Versus best verified row Best verified: Claude Sonnet 5.5 · 69.2% | Gap21.1 behind | WeightDisplay only | Benchmark exact |
Math2 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| FrontierMath v2 (Tiers 1-3)FrontierMath v2 Tiers 1-3 | Score78.596% | Versus best verified row Best verified: GPT-6 Astra · 93.684% | Gap15.1 behind | WeightWeighted 30% | Benchmark exact |
| FrontierMath v2 (Tier 4)FrontierMath v2 Tier 4 | Score49.000% | Versus best verified row Best verified: GPT-6.1 Sol · 100.000% | Gap51 behind | WeightWeighted 10% | Benchmark exact |
Bars run 0–100; the dark tick marks the best source-verified value
All 42 rowsWhat it costs to get this score
Published API price against the public score. The x-axis uses a log scale; the dashed path marks models that are not beaten by a cheaper, higher-scoring option. Price uses average of published input and output rates.
GPT-5.4 · 66.7 score · $8.75 blended per million tokens
The chart opens on the current model. Scroll horizontally to inspect the full price axis.
Horizontal: blended price per million tokens, log scale · Vertical: public score
Lineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- gpt-5.4OpenAI pricing
- Context window
- 1.05MOpenAI pricing
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- Not sourced yet
- Output modalities
- Not sourced yet
- Parameters
- Not disclosed by the provider
- Availability
- Not sourced yet
- Cloud regions
- Not tracked yet
- Lifecycle
- Superseded
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Published at $0.25 per million cached input tokensOpenAI pricing
- Self-host
- Weights are not published
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
GPT-5.4 ranks #30 of 224 on the public leaderboard with a score of 66.7/100. Its source-verified position is #21 of 88.
GPT-5.4 is a proprietary model with a 1.05M context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
ExploitGym is normalized from 54 successful exploits out of 898 tasks in the public paper. HealthBench Professional uses the base GPT-5.4 overall score from the public HealthBench Professional paper, not the ChatGPT for Clinicians harness row.
GPT-5.4 sits in the GPT-5.4 family with GPT-5.4 mini, GPT-5.4 nano, GPT-5.4 Pro. 42 of 681 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Knowledge at #20, while its lowest eligible position is Coding at #48. particularly effective for knowledge-intensive tasks like research, analysis, and factual Q&A.
Last updated October 17, 2026. Runtime fields remain blank until a sourced snapshot exists.
Questions
How does GPT-5.4 perform overall in AI benchmarks?
GPT-5.4 ranks #30 out of 224 models on the public BenchAlign leaderboard, with a score of 66.7/100. Its evidence status is Supported, and this profile shows 42 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
Is GPT-5.4 good for knowledge and understanding?
GPT-5.4 ranks #20 out of 184 eligible models for knowledge and understanding, with a public category score of 69.1/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is GPT-5.4 good for coding and programming?
GPT-5.4 ranks #48 out of 150 eligible models for coding and programming, with a public category score of 47.7/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is GPT-5.4 good for mathematics?
GPT-5.4 has source-displayable benchmark coverage for mathematics, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is GPT-5.4 good for reasoning and logic?
GPT-5.4 ranks #23 out of 28 eligible models for reasoning and logic, with a public category score of 64.2/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is GPT-5.4 good for agentic tool use and computer tasks?
GPT-5.4 ranks #44 out of 128 eligible models for agentic tool use and computer tasks, with a public category score of 52.1/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is GPT-5.4 good for multimodal and grounded tasks?
GPT-5.4 ranks #30 out of 54 eligible models for multimodal and grounded tasks, with a public category score of 73.1/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Which sibling models are related to GPT-5.4?
GPT-5.4 belongs to the GPT-5.4 family. Related tracked variants include GPT-5.4 mini, GPT-5.4 nano, GPT-5.4 Pro. A sibling link indicates shared lineage or a documented configuration relationship; it does not mean the variants have identical pricing, context limits, benchmark evidence, or deployment behavior. Compare before switching.
Does GPT-5.4 have full benchmark coverage on BenchLM?
No. GPT-5.4 currently has 57 source-displayable rows across 681 tracked benchmark slots. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of GPT-5.4?
GPT-5.4 has a documented context window of 1.05M. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.