# GPT-5.4 Pro vs GPT-5.5

> Side-by-side benchmark comparison for GPT-5.4 Pro and GPT-5.5 across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/gpt-5-4-pro-vs-gpt-5-5
- Last updated: September 4, 2026

- Shared sourced benchmarks: 7
- HTML indexing: indexable
- Ranking lane: BenchAlign v5

## Quick Verdict

Pick GPT-5.5 if you want the stronger benchmark profile. GPT-5.4 Pro only makes more sense when its price, context window, or workload-specific category wins matter more than the overall score.

## Summary

- GPT-5.5 leads overall 73.27 to 61.75.
- The clearest category separation is in knowledge, where the averages are 60.7 for GPT-5.4 Pro and 73.3 for GPT-5.5.
- The biggest single benchmark swing is HLE in Knowledge, with scores of 58.7% and 52.2%.
- GPT-5.5 is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- GPT-5.4 Pro also has the larger context window at 1.05M.

## Model Snapshot

| Property | GPT-5.4 Pro | GPT-5.5 |
|----------|----------|----------|
| Creator | OpenAI | OpenAI |
| Type | Proprietary | Proprietary |
| Reasoning | Reasoning | Reasoning |
| Context | 1.05M | 1M |
| Overall Score | 61.75 | 73.27 |
| Benchmarks Covered | 10 | 38 |
| Pricing (input/output) | $30.00 / $180.00 | $5.00 / $30.00 |

## Category Breakdown

### Agentic

- Winner: Coming soon
- GPT-5.4 Pro public-lane score: 58.1 (Estimated · #33/151)
- GPT-5.5 public-lane score: 63.9 (Supported · #15/151)

| Benchmark | GPT-5.4 Pro | GPT-5.5 | Winner |
|-----------|-----------|-----------|--------|
| BrowseComp | 89.3% | 84.4% | GPT-5.4 Pro |
| Terminal-Bench 2.0 | Coming soon | 82% | Coming soon |
| CyberGym | Coming soon | 81.8% | Coming soon |
| OSWorld-Verified | Coming soon | 78.7% | Coming soon |
| MCP Atlas | Coming soon | 75.3% | Coming soon |
| Toolathlon | Coming soon | 55.6% | Coming soon |
| τ²-bench results | Coming soon | 98% | Coming soon |
| Gert Labs | Coming soon | 72.93% | Coming soon |
| ResearchClawBench | Coming soon | 17.0% | Coming soon |
| OSWorld 2.0 | Coming soon | 13.0% | Coming soon |
| JobBench | Coming soon | 42.7% | Coming soon |
| ExploitGym | Coming soon | 13.4% | Coming soon |
| Terminal-Bench 2.1 (Vals) | Coming soon | 76.4% | Coming soon |

### Coding

- Winner: Coming soon
- GPT-5.4 Pro public-lane score: Coming soon
- GPT-5.5 public-lane score: 67.7 (Supported · #8/183)

| Benchmark | GPT-5.4 Pro | GPT-5.5 | Winner |
|-----------|-----------|-----------|--------|
| SWE-bench Pro | Coming soon | 58.6% | Coming soon |
| Terminal-Bench 2.0 | Coming soon | 82.0% | Coming soon |
| Vibe Code Bench | Coming soon | 69.85% | Coming soon |
| React Native Evals | Coming soon | 84.7% | Coming soon |
| cursorBench31 | Coming soon | 59.2% | Coming soon |
| cursorBench32 | Coming soon | 58.4% | Coming soon |
| FrontierCode 1.1 Main | Coming soon | 43.0% | Coming soon |
| LiveCodeBench (Vals) | Coming soon | 85.3% | Coming soon |
| SWE-bench (Vals) | Coming soon | 82.6% | Coming soon |

### Multimodal & Grounded

- Winner: Coming soon
- GPT-5.4 Pro public-lane score: Coming soon
- GPT-5.5 public-lane score: 71.3 (#19/48)

| Benchmark | GPT-5.4 Pro | GPT-5.5 | Winner |
|-----------|-----------|-----------|--------|
| MMMU-Pro | Coming soon | 81.2% | Coming soon |
| MMMU-Pro w/ Python | Coming soon | 83.2% | Coming soon |
| OfficeQA Pro | Coming soon | 54.1% | Coming soon |

### Reasoning

- Winner: Coming soon
- GPT-5.4 Pro public-lane score: 70.1 (Unranked · 2 rankable rows)
- GPT-5.5 public-lane score: 63.5 (#15/22)

| Benchmark | GPT-5.4 Pro | GPT-5.5 | Winner |
|-----------|-----------|-----------|--------|
| ARC-AGI-2 | 83.3% | 85% | GPT-5.5 |
| MRCR v2 64K-128K | Coming soon | 83.1% | Coming soon |
| MRCR v2 128K-256K | Coming soon | 87.5% | Coming soon |
| ARC-AGI-3 | Coming soon | 0.4% | Coming soon |

### Knowledge

- Winner: Coming soon
- GPT-5.4 Pro public-lane score: 60.7 (Estimated · #39/181)
- GPT-5.5 public-lane score: 73.3 (Supported · #7/181)

| Benchmark | GPT-5.4 Pro | GPT-5.5 | Winner |
|-----------|-----------|-----------|--------|
| HLE | 58.7% | 52.2% | GPT-5.4 Pro |
| FrontierScience | 36.7% | Coming soon | Coming soon |
| FrontierScience Research | 36.7% | Coming soon | Coming soon |
| HLE w/o tools | 42.7% | 41.4% | GPT-5.4 Pro |
| GPQA | Coming soon | 93.6% | Coming soon |
| GPQA-D | Coming soon | 93.6% | Coming soon |
| GPQA Diamond (Vals) | Coming soon | 93.2% | Coming soon |
| MMLU-Pro (Vals) | Coming soon | 88.1% | Coming soon |

### Instruction Following

- Winner: Coming soon
- GPT-5.4 Pro public-lane score: Coming soon
- GPT-5.5 public-lane score: 92.9 (#7/120)

### Mathematics

- Winner: Coming soon
- GPT-5.4 Pro public-lane score: 69 (Unranked · 4 rankable rows)
- GPT-5.5 public-lane score: 69.6 (Unranked · 3 rankable rows)

| Benchmark | GPT-5.4 Pro | GPT-5.5 | Winner |
|-----------|-----------|-----------|--------|
| IPhO 2025 (Theory) | 93.5% | Coming soon | Coming soon |
| FrontierMath (legacy) | 50% | 51.7% | GPT-5.5 |
| FrontierMath v2 (Tiers 1-3) | 50.000% | 51.700% | GPT-5.5 |
| FrontierMath v2 (Tier 4) | 37.500% | 35.400% | GPT-5.4 Pro |

## FAQ

### Which is better overall, GPT-5.4 Pro or GPT-5.5?

GPT-5.5 is ahead overall on BenchLM right now.

### Where is the biggest gap between GPT-5.4 Pro and GPT-5.5?

The widest category gap is in knowledge, where the averages are 60.7 for GPT-5.4 Pro and 73.3 for GPT-5.5.

### How many benchmarks does GPT-5.4 Pro cover on BenchLM?

GPT-5.4 Pro currently has 10 sourced benchmark scores on BenchLM.

### How many benchmarks does GPT-5.5 cover on BenchLM?

GPT-5.5 currently has 38 sourced benchmark scores on BenchLM.

## Related Comparisons

- [GPT-5.4 Pro vs GPT-5.4](/compare/gpt-5-4-vs-gpt-5-4-pro)
- [GPT-5.5 vs GPT-5.4](/compare/gpt-5-4-vs-gpt-5-5)
- [GPT-5.4 Pro vs GPT-5.4 mini](/compare/gpt-5-4-mini-vs-gpt-5-4-pro)
- [GPT-5.5 vs GPT-5.4 mini](/compare/gpt-5-4-mini-vs-gpt-5-5)

## Explore More

- [GPT-5.4 Pro profile](/models/gpt-5-4-pro)
- [GPT-5.5 profile](/models/gpt-5-5)
- [Compare Pricing](/llm-pricing)
- [Alternative Finder](/tools/alternative-finder)
- [LLM Selector](/tools/llm-selector)
- [Overall Rankings](/best/overall)