Model comparison
GPT-5.4 vs GPT-5.4 nano
Head-to-head evidence from 29 shared benchmark results across 7 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Sibling matchup inside the GPT-5.4 family.
Public leaderboard positions: GPT-5.4 #8 (Supported); GPT-5.4 nano #25 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.4 and GPT-5.4 nano share 29 comparable benchmark results. 4 of 8 categories are comparable. 23 results are unique to GPT-5.4; 0 to GPT-5.4 nano.
Updated July 23, 2026- Shared results
- 29
- GPT-5.4 only
- 23
- GPT-5.4 nano only
- 0
- Comparable categories
- 4 / 8
GPT-5.4 makes more sense if agentic is the priority or you need the larger 1.05M context window, while GPT-5.4 nano is the cleaner fit if you want the cheaper token bill.
Confidence note. This is a partial-evidence comparison with 29 shared benchmark results across 7 evidence categories; 4 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
GPT-5.4 and GPT-5.4 nano sit in the same GPT-5.4 family. This page is less about two unrelated model lineages and more about how the siblings trade off on benchmark shape, token costs, and practical limits like context window.
GPT-5.4 is clearly ahead on the BenchAlign aggregate, 74.24 to 66.79. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.4's sharpest advantage is in agentic, where it averages 77.2 against 42.9. The single biggest benchmark swing on the page is OSWorld-Verified, 75% to 39%.
GPT-5.4 is also the more expensive model on tokens at $2.50 input / $15.00 output per 1M tokens, versus $0.20 input / $1.25 output per 1M tokens for GPT-5.4 nano. That is roughly 12.0x on output cost alone. GPT-5.4 gives you the larger context window at 1.05M, compared with 400K for GPT-5.4 nano.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | GPT-5.4 | Δ | GPT-5.4 nano |
|---|---|---|---|
| Agentic | GPT-5.477.2 | Margin← 34.3 | GPT-5.4 nano42.9 |
| Math | GPT-5.442.5 | Margin← 21.5 | GPT-5.4 nano21.0 |
| Knowledge | GPT-5.457.6 | Margin← 13.8 | GPT-5.4 nano43.8 |
| Multimodal | GPT-5.473.2 | Margin← 7.1 | GPT-5.4 nano66.1 |
| Coding | GPT-5.457.7 | MarginNo overlap | GPT-5.4 nanoNot measured |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
OSWorld-Verified
AgenticA 75%B 39%Winner: GPT-5.4Δ 36OSWorld-Verified: GPT-5.4 scored 75%; GPT-5.4 nano scored 39%. GPT-5.4 wins this benchmark. - Source ↗
Terminal-Bench 2.0
AgenticA 75.1%B 46.3%Winner: GPT-5.4Δ 28.8Terminal-Bench 2.0: GPT-5.4 scored 75.1%; GPT-5.4 nano scored 46.3%. GPT-5.4 wins this benchmark. - Source ↗
FrontierMath v2 (Tiers 1-3)
MathA 47.600%B 25.860%Winner: GPT-5.4Δ 21.7FrontierMath v2 (Tiers 1-3): GPT-5.4 scored 47.600%; GPT-5.4 nano scored 25.860%. GPT-5.4 wins this benchmark. - Source ↗
FrontierMath v2 (Tier 4)
MathA 27.100%B 6.250%Winner: GPT-5.4Δ 20.9FrontierMath v2 (Tier 4): GPT-5.4 scored 27.100%; GPT-5.4 nano scored 6.250%. GPT-5.4 wins this benchmark. - Source ↗
MMMU-Pro
MultimodalA 81.2%B 66.1%Winner: GPT-5.4Δ 15.1MMMU-Pro: GPT-5.4 scored 81.2%; GPT-5.4 nano scored 66.1%. GPT-5.4 wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.4 | GPT-5.4 nano | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.4$2.5 input / $15 output | GPT-5.4 nano$0.2 input / $1.25 output | GPT-5.4 nano has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.474 tok/s | GPT-5.4 nano191 tok/s | GPT-5.4 nano has the higher measured throughput. |
| First-answer latencyseconds to first token | GPT-5.4151.79 s | GPT-5.4 nano3.64 s | GPT-5.4 nano reaches the first token sooner. |
| Context windowmaximum listed tokens | GPT-5.41.05M | GPT-5.4 nano400K | GPT-5.4 lists the larger context window. |
Benchmark Deep Dive
AgenticGPT-5.4 wins17 benchmarks
| Benchmark | GPT-5.4 | GPT-5.4 nano | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 75.1% | 46.3% | GPT-5.4 leads |
| CyberGymSource | 79.0% | — | Not comparable |
| BrowseCompSource | 82.7% | — | Not comparable |
| OSWorld-VerifiedSource | 75% | 39% | GPT-5.4 leads |
| MCP AtlasSource | 70.6% | 56.1% | GPT-5.4 leads |
| ToolathlonSource | 54.6% | 35.5% | GPT-5.4 leads |
| τ²-bench resultsSource | 87.1% | 76% | GPT-5.4 leads |
| Claw-EvalSource | 60.3% | — | Not comparable |
| DeepSearchQASource | 73.6% | — | Not comparable |
| AA Agentic IndexSource | 41.1% | 27.5% | GPT-5.4 leads |
| APEX-Agents-AASource | 33.3% | 24.9% | GPT-5.4 leads |
| GDPval-AASource | 44.7% | 30.0% | GPT-5.4 leads |
| GDPval-AASource | 1395 | 1100 | GPT-5.4 leads |
| Gert LabsSource | 64.89% | — | Not comparable |
| ResearchClawBenchSource | 15.3% | — | Not comparable |
| JobBenchSource | 38.9% | — | Not comparable |
| ExploitGymSource | 6.0% | — | Not comparable |
Coding6 benchmarks
Reasoning2 benchmarks
KnowledgeGPT-5.4 wins13 benchmarks
| Benchmark | GPT-5.4 | GPT-5.4 nano | Result |
|---|---|---|---|
| GPQASource | 92.8% | 82.8% | GPT-5.4 leads |
| HLESource | 52.1% | 37.7% | GPT-5.4 leads |
| HLE w/o toolsSource | 39.8% | 24.3% | GPT-5.4 leads |
| GPQA-DSource | 92.8% | — | Not comparable |
| HealthBench HardSource | 40.1% | — | Not comparable |
| MedXpertQA (Text)Source | 59.6% | — | Not comparable |
| Artificial Analysis Intelligence IndexSource | 51.4% | 38.2% | GPT-5.4 leads |
| AA-GPQA DiamondSource | 92.0% | 81.7% | GPT-5.4 leads |
| AA-HLESource | 41.6% | 26.5% | GPT-5.4 leads |
| AA-Omniscience IndexSource | 5.7% | -29.5% | GPT-5.4 leads |
| AA-Omniscience AccuracySource | 50.0% | 25.4% | GPT-5.4 leads |
| AA-Omniscience Hallucination RateSource | 88.6% | 73.6% | GPT-5.4 nano leads |
| HealthBench ProfessionalSource | 48.1% | — | Not comparable |
MathGPT-5.4 wins2 benchmarks
MultimodalGPT-5.4 wins11 benchmarks
| Benchmark | GPT-5.4 | GPT-5.4 nano | Result |
|---|---|---|---|
| MMMU-ProSource | 81.2% | 66.1% | GPT-5.4 leads |
| OfficeQA ProSource | 53.2% | — | Not comparable |
| MMMU-Pro w/ PythonSource | 82.1% | 69.5% | GPT-5.4 leads |
| CharXivSource | 82.8% | — | Not comparable |
| ERQASource | 65.4% | — | Not comparable |
| SimpleVQASource | 61.1% | — | Not comparable |
| ScreenSpot ProSource | 85.4% | — | Not comparable |
| ZeroBenchSource | 41.0% | — | Not comparable |
| MedXpertQA (MM)Source | 77.1% | — | Not comparable |
| AA-MMMU-ProSource | 78.4% | 65.4% | GPT-5.4 leads |
| Design Arena WebsiteSource | 1250 | — | Not comparable |
Inst. Following1 benchmarks
| Benchmark | GPT-5.4 | GPT-5.4 nano | Result |
|---|---|---|---|
| AA-IFBenchSource | 73.9% | 75.9% | GPT-5.4 nano leads |
Frequently Asked Questions (5)
Which is better, GPT-5.4 or GPT-5.4 nano?
GPT-5.4 and GPT-5.4 nano are sibling variants in the GPT-5.4 family, so the right pick depends on whether you value the better benchmark line, cheaper tokens, or the larger context window. GPT-5.4 is ahead on BenchLM's BenchAlign leaderboard 74.24 to 66.79.
Which is better for knowledge tasks, GPT-5.4 or GPT-5.4 nano?
GPT-5.4 has the edge for knowledge tasks in this comparison, averaging 57.6 versus 43.8. Inside this category, AA-Omniscience Index is the benchmark that creates the most daylight between them.
Which is better for math, GPT-5.4 or GPT-5.4 nano?
GPT-5.4 has the edge for math in this comparison, averaging 42.5 versus 21. Inside this category, FrontierMath v2 (Tiers 1-3) is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, GPT-5.4 or GPT-5.4 nano?
GPT-5.4 has the edge for agentic tasks in this comparison, averaging 77.2 versus 42.9. Inside this category, GDPval-AA is the benchmark that creates the most daylight between them.
Which is better for multimodal and grounded tasks, GPT-5.4 or GPT-5.4 nano?
GPT-5.4 has the edge for multimodal and grounded tasks in this comparison, averaging 73.2 versus 66.1. Inside this category, MMMU-Pro is the benchmark that creates the most daylight between them.
Related Comparisons
Explore More
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.