# GPT-5 vs GPT-5.4

> Side-by-side benchmark comparison for GPT-5 and GPT-5.4 across agentic, coding, multimodal, knowledge, reasoning, multilingual, and math tasks.

- Canonical page: https://benchlm.ai/compare/gpt-5-vs-gpt-5-4
- Last updated: November 18, 2026

- Shared sourced benchmarks: 6
- HTML indexing: indexable
- Ranking lane: BenchAlign v5.8

## Quick Verdict

GPT-5.4 leads by point estimate. Conditional score ranges do not establish rank confidence. Choose using the category evidence and your own workload tests.

## Summary

- GPT-5.4 has the higher overall point estimate 64.12 to 48.96. Conditional score ranges do not establish rank confidence.
- The clearest category separation is in reasoning, where the averages are 43.7 for GPT-5 and 61.8 for GPT-5.4.
- The biggest single benchmark swing is Vibe Code Bench in Coding, with scores of 20.09% and 67.42%.
- GPT-5 is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.
- GPT-5.4 also has the larger context window at 1.05M.

## Model Snapshot

| Property | GPT-5 | GPT-5.4 |
|----------|----------|----------|
| Creator | OpenAI | OpenAI |
| Type | Proprietary | Proprietary |
| Reasoning | Reasoning | Reasoning |
| Context | 128K | 1.05M |
| Overall Score | 48.96 | 64.12 |
| Benchmarks Covered | 6 | 51 |
| Pricing (input/output) | $1.25 / $10.00 | $2.50 / $15.00 |

## Category Breakdown

### Agentic

- Winner: Not comparable
- GPT-5 public-lane score: Coming soon
- GPT-5.4 public-lane score: 51.6 (Supported · #57/138)

| Benchmark | GPT-5 | GPT-5.4 | Winner |
|-----------|-----------|-----------|--------|
| JobBench | 8.5% | 38.9% | GPT-5.4 |
| LifeVal | Coming soon | 73.0% | Coming soon |
| Terminal-Bench 2.0 | Coming soon | 75.1% | Coming soon |
| CyberGym | Coming soon | 79.0% | Coming soon |
| BrowseComp | Coming soon | 82.7% | Coming soon |
| OSWorld-Verified | Coming soon | 75% | Coming soon |
| MCP Atlas | Coming soon | 70.6% | Coming soon |
| Toolathlon | Coming soon | 54.6% | Coming soon |
| τ²-bench results | Coming soon | 98.9% | Coming soon |
| Claw-Eval | Coming soon | 60.3% | Coming soon |
| DeepSearchQA | Coming soon | 73.6% | Coming soon |
| Gert Labs | Coming soon | 64.89% | Coming soon |
| ResearchClawBench | Coming soon | 15.3% | Coming soon |
| ExploitGym | Coming soon | 6.0% | Coming soon |
| ApprenticeBench | Coming soon | 11% | Coming soon |
| MazeBench | Coming soon | 0% | Coming soon |

### Coding

- Winner: Directional only
- GPT-5 public-lane score: 36.7 (Estimated · #85/160)
- GPT-5.4 public-lane score: 48.8 (Supported · #50/160)

| Benchmark | GPT-5 | GPT-5.4 | Winner |
|-----------|-----------|-----------|--------|
| Vibe Code Bench | 20.09% | 67.42% | GPT-5.4 |
| LiveCodeBench (Vals) | 85.9% | 84.1% | GPT-5 |
| SWE-bench (Vals) | 69.0% | 78.2% | GPT-5.4 |
| LiveCodeBench Pro | Coming soon | 87.5% | Coming soon |
| SWE-bench Pro | Coming soon | 57.7% | Coming soon |
| React Native Evals | Coming soon | 85.3% | Coming soon |
| Terminal-Bench 2.0 | Coming soon | 75.1% | Coming soon |
| PostTrainBench v1.1 | Coming soon | 19.0% | Coming soon |
| PostTrainBench | Coming soon | 17.8% | Coming soon |
| PostTrainBench v1 | Coming soon | 20.2% | Coming soon |

### Reasoning

- Winner: Directional only
- GPT-5 public-lane score: 43.7 (Estimated · #75/159)
- GPT-5.4 public-lane score: 61.8 (Supported · #27/159)

| Benchmark | GPT-5 | GPT-5.4 | Winner |
|-----------|-----------|-----------|--------|
| BioBasicsBench | Coming soon | 33.2% | Coming soon |
| ARC-AGI-2 | Coming soon | 74.0% | Coming soon |
| ARC-AGI-3 | Coming soon | 0.2% | Coming soon |
| ARC-AGI-1 | Coming soon | 93.67% | Coming soon |

### Multimodal & Grounded

- Winner: Not comparable
- GPT-5 public-lane score: Coming soon
- GPT-5.4 public-lane score: 63.3 (Estimated · #29/66)

| Benchmark | GPT-5 | GPT-5.4 | Winner |
|-----------|-----------|-----------|--------|
| ParseBench-1 | Coming soon | 61.72% | Coming soon |
| MMMU-Pro | Coming soon | 81.2% | Coming soon |
| OfficeQA Pro | Coming soon | 53.2% | Coming soon |
| MMMU-Pro w/ Python | Coming soon | 82.1% | Coming soon |
| CharXiv | Coming soon | 82.8% | Coming soon |
| ERQA | Coming soon | 65.4% | Coming soon |
| SimpleVQA | Coming soon | 61.1% | Coming soon |
| ScreenSpot Pro | Coming soon | 85.4% | Coming soon |
| ZeroBench | Coming soon | 41.0% | Coming soon |
| MedXpertQA (MM) | Coming soon | 77.1% | Coming soon |

### Knowledge

- Winner: GPT-5.4
- GPT-5 public-lane score: 53.1 (Supported · #67/179)
- GPT-5.4 public-lane score: 65.6 (Supported · #28/179)

| Benchmark | GPT-5 | GPT-5.4 | Winner |
|-----------|-----------|-----------|--------|
| GPQA Diamond (Vals) | 85.6% | 91.7% | GPT-5.4 |
| MMLU-Pro (Vals) | 86.5% | 87.5% | GPT-5.4 |
| GPQA | Coming soon | 92.8% | Coming soon |
| HLE | Coming soon | 52.1% | Coming soon |
| HLE w/o tools | Coming soon | 39.8% | Coming soon |
| GPQA-D | Coming soon | 92.8% | Coming soon |
| HealthBench Hard | Coming soon | 40.1% | Coming soon |
| MedXpertQA (Text) | Coming soon | 59.6% | Coming soon |
| HealthBench Professional | Coming soon | 48.1% | Coming soon |

### Instruction Following

- Winner: Directional only
- GPT-5 public-lane score: 89.4 (#35/157)
- GPT-5.4 public-lane score: 90.4 (#27/157)

### Mathematics

- Winner: Not comparable
- GPT-5 public-lane score: Coming soon
- GPT-5.4 public-lane score: 75.5 (Unranked · 2 rankable rows)

| Benchmark | GPT-5 | GPT-5.4 | Winner |
|-----------|-----------|-----------|--------|
| FrontierMath v2 (Tiers 1-3) | Coming soon | 78.596% | Coming soon |
| FrontierMath v2 (Tier 4) | Coming soon | 49.000% | Coming soon |

## FAQ

### Which is better overall, GPT-5 or GPT-5.4?

GPT-5.4 has the higher overall point estimate. Conditional score ranges do not establish rank confidence.

### Where is the biggest gap between GPT-5 and GPT-5.4?

The widest category gap is in reasoning, where the averages are 43.7 for GPT-5 and 61.8 for GPT-5.4.

### How many benchmarks does GPT-5 cover on BenchLM?

GPT-5 currently has 6 sourced benchmark scores on BenchLM.

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

GPT-5.4 currently has 51 sourced benchmark scores on BenchLM.

## Related Comparisons

- [GPT-5 vs GPT-5 mini](/compare/gpt-5-vs-gpt-5-mini)
- [GPT-5.4 vs GPT-5 mini](/compare/gpt-5-4-vs-gpt-5-mini)
- [GPT-5 vs GPT-5 nano](/compare/gpt-5-vs-gpt-5-nano)
- [GPT-5.4 vs GPT-5 nano](/compare/gpt-5-4-vs-gpt-5-nano)

## Explore More

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