Model comparison
GPT-5.4 mini vs GPT-5.5 Pro
Head-to-head evidence from 5 shared benchmark results across 3 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: GPT-5.4 mini #75 (Estimated); GPT-5.5 Pro #38 (Estimated). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. GPT-5.4 mini and GPT-5.5 Pro share 5 comparable benchmark results. 3 of 8 categories are comparable. 25 results are unique to GPT-5.4 mini; 2 to GPT-5.5 Pro.
Updated July 22, 2026- Shared results
- 5
- GPT-5.4 mini only
- 25
- GPT-5.5 Pro only
- 2
- Comparable categories
- 3 / 8
Pick GPT-5.5 Pro if you want the stronger benchmark profile. GPT-5.4 mini only becomes the better choice if you want the cheaper token bill.
Confidence note. This is a partial-evidence comparison with 5 shared benchmark results across 3 evidence categories; 3 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
GPT-5.5 Pro is clearly ahead on the BenchAlign aggregate, 63.69 to 56.77. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.
GPT-5.5 Pro's sharpest advantage is in mathematics, where it averages 48.1 against 21.7. The single biggest benchmark swing on the page is FrontierMath v2 (Tier 4), 2.080% to 39.600%.
GPT-5.5 Pro is also the more expensive model on tokens at $30.00 input / $180.00 output per 1M tokens, versus $0.75 input / $4.50 output per 1M tokens for GPT-5.4 mini. That is roughly 40.0x on output cost alone. GPT-5.5 Pro gives you the larger context window at 1M, compared with 400K for GPT-5.4 mini.
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 mini | Δ | GPT-5.5 Pro |
|---|---|---|---|
| Math | GPT-5.4 mini21.7 | Margin→ 26.4 | GPT-5.5 Pro48.1 |
| Agentic | GPT-5.4 mini65.7 | Margin→ 24.4 | GPT-5.5 Pro90.1 |
| Knowledge | GPT-5.4 mini47.8 | Margin→ 9.4 | GPT-5.5 Pro57.2 |
| Multimodal | GPT-5.4 mini76.6 | MarginNo overlap | GPT-5.5 ProNot measured |
Decisive benchmark drivers
The largest measured benchmark gaps in this matchup, with exact reported values.
More
- Source ↗
FrontierMath v2 (Tier 4)
MathA 2.080%B 39.600%Winner: GPT-5.5 ProΔ 37.5FrontierMath v2 (Tier 4): GPT-5.4 mini scored 2.080%; GPT-5.5 Pro scored 39.600%. GPT-5.5 Pro wins this benchmark. - Source ↗
FrontierMath v2 (Tiers 1-3)
MathA 28.280%B 51.000%Winner: GPT-5.5 ProΔ 22.7FrontierMath v2 (Tiers 1-3): GPT-5.4 mini scored 28.280%; GPT-5.5 Pro scored 51.000%. GPT-5.5 Pro wins this benchmark. - Source ↗
HLE
KnowledgeA 41.5%B 57.2%Winner: GPT-5.5 ProΔ 15.7HLE: GPT-5.4 mini scored 41.5%; GPT-5.5 Pro scored 57.2%. GPT-5.5 Pro wins this benchmark.
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | GPT-5.4 mini | GPT-5.5 Pro | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | GPT-5.4 mini$0.75 input / $4.5 output | GPT-5.5 Pro$30 input / $180 output | GPT-5.4 mini has the lower combined listed price. |
| Generation speedtokens per second | GPT-5.4 mini201 tok/s | GPT-5.5 ProNot available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | GPT-5.4 mini3.85 s | GPT-5.5 ProNot available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | GPT-5.4 mini400K | GPT-5.5 Pro1M | GPT-5.5 Pro lists the larger context window. |
Benchmark Deep Dive
AgenticGPT-5.5 Pro wins10 benchmarks
| Benchmark | GPT-5.4 mini | GPT-5.5 Pro | Result |
|---|---|---|---|
| Terminal-Bench 2.0Source | 60% | — | Not comparable |
| OSWorld-VerifiedSource | 72.1% | — | Not comparable |
| MCP AtlasSource | 57.7% | — | Not comparable |
| ToolathlonSource | 42.9% | — | Not comparable |
| τ²-bench resultsSource | 83.3% | — | Not comparable |
| AA Agentic IndexSource | 30.2% | — | Not comparable |
| APEX-Agents-AASource | 28.2% | — | Not comparable |
| GDPval-AASource | 33.6% | — | Not comparable |
| GDPval-AASource | 1171 | — | Not comparable |
| BrowseCompSource | — | 90.1% | Not comparable |
Coding4 benchmarks
Reasoning2 benchmarks
KnowledgeGPT-5.5 Pro wins9 benchmarks
| Benchmark | GPT-5.4 mini | GPT-5.5 Pro | Result |
|---|---|---|---|
| GPQASource | 88% | — | Not comparable |
| HLESource | 41.5% | 57.2% | GPT-5.5 Pro leads |
| HLE w/o toolsSource | 28.2% | 43.1% | GPT-5.5 Pro leads |
| Artificial Analysis Intelligence IndexSource | 40.0% | — | Not comparable |
| AA-GPQA DiamondSource | 87.5% | — | Not comparable |
| AA-HLESource | 26.6% | — | Not comparable |
| AA-Omniscience IndexSource | -18.7% | — | Not comparable |
| AA-Omniscience AccuracySource | 37.5% | — | Not comparable |
| AA-Omniscience Hallucination RateSource | 89.8% | — | Not comparable |
MathGPT-5.5 Pro wins3 benchmarks
Multimodal3 benchmarks
Inst. Following1 benchmarks
| Benchmark | GPT-5.4 mini | GPT-5.5 Pro | Result |
|---|---|---|---|
| AA-IFBenchSource | 73.3% | — | Not comparable |
Frequently Asked Questions (4)
Which is better, GPT-5.4 mini or GPT-5.5 Pro?
GPT-5.5 Pro is ahead on BenchLM's BenchAlign leaderboard, 63.69 to 56.77. The biggest single separator in this matchup is FrontierMath v2 (Tier 4), where the scores are 2.080% and 39.600%.
Which is better for knowledge tasks, GPT-5.4 mini or GPT-5.5 Pro?
GPT-5.5 Pro has the edge for knowledge tasks in this comparison, averaging 57.2 versus 47.8. Inside this category, HLE is the benchmark that creates the most daylight between them.
Which is better for math, GPT-5.4 mini or GPT-5.5 Pro?
GPT-5.5 Pro has the edge for math in this comparison, averaging 48.1 versus 21.7. Inside this category, FrontierMath v2 (Tier 4) is the benchmark that creates the most daylight between them.
Which is better for agentic tasks, GPT-5.4 mini or GPT-5.5 Pro?
GPT-5.5 Pro has the edge for agentic tasks in this comparison, averaging 90.1 versus 65.7. GPT-5.4 mini stays close enough that the answer can still flip depending on your workload.
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.