# GPT-5.4 mini vs GPT-5.4 nano

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

- Canonical page: https://benchlm.ai/compare/gpt-5-4-mini-vs-gpt-5-4-nano
- Last updated: September 10, 2026

- Shared sourced benchmarks: 18
- HTML indexing: indexable
- Ranking lane: BenchAlign v5

## Quick Verdict

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

## Summary

- GPT-5.4 mini leads overall 61.11 to 59.57.
- The clearest category separation is in multimodal & grounded, where the averages are 57.2 for GPT-5.4 mini and 23.8 for GPT-5.4 nano.
- The biggest single benchmark swing is OSWorld-Verified in Agentic, with scores of 72.1% and 39%.
- GPT-5.4 nano is the cheaper option on output tokens, which matters if you expect large responses or heavy interactive use.

## Model Snapshot

| Property | GPT-5.4 mini | GPT-5.4 nano |
|----------|----------|----------|
| Creator | OpenAI | OpenAI |
| Type | Proprietary | Proprietary |
| Reasoning | Reasoning | Reasoning |
| Context | 400K | 400K |
| Overall Score | 61.11 | 59.57 |
| Benchmarks Covered | 19 | 18 |
| Pricing (input/output) | $0.75 / $4.50 | $0.20 / $1.25 |

## Category Breakdown

### Agentic

- Winner: Coming soon
- GPT-5.4 mini public-lane score: 39.1 (Estimated · #119/152)
- GPT-5.4 nano public-lane score: 34.6 (Supported · #133/152)

| Benchmark | GPT-5.4 mini | GPT-5.4 nano | Winner |
|-----------|-----------|-----------|--------|
| Terminal-Bench 2.0 | 60% | 46.3% | GPT-5.4 mini |
| OSWorld-Verified | 72.1% | 39% | GPT-5.4 mini |
| MCP Atlas | 57.7% | 56.1% | GPT-5.4 mini |
| Toolathlon | 42.9% | 35.5% | GPT-5.4 mini |
| τ²-bench results | 93.4% | 92.5% | GPT-5.4 mini |
| Terminal-Bench 2.1 (Vals) | 54.7% | 41.6% | GPT-5.4 mini |

### Coding

- Winner: GPT-5.4 mini
- GPT-5.4 mini public-lane score: 42.7 (Supported · #104/151)
- GPT-5.4 nano public-lane score: 37.1 (Supported · #126/151)

| Benchmark | GPT-5.4 mini | GPT-5.4 nano | Winner |
|-----------|-----------|-----------|--------|
| Vibe Code Bench | 47.97% | 26.10% | GPT-5.4 mini |
| FrontierCode 1.1 Main | 27.0% | Coming soon | Coming soon |
| LiveCodeBench (Vals) | 81.5% | 84.0% | GPT-5.4 nano |
| SWE-bench (Vals) | 73.0% | 69.8% | GPT-5.4 mini |

### Multimodal & Grounded

- Winner: GPT-5.4 mini
- GPT-5.4 mini public-lane score: 57.2 (#31/48)
- GPT-5.4 nano public-lane score: 23.8 (#45/48)

| Benchmark | GPT-5.4 mini | GPT-5.4 nano | Winner |
|-----------|-----------|-----------|--------|
| MMMU-Pro | 76.6% | 66.1% | GPT-5.4 mini |
| MMMU-Pro w/ Python | 78% | 69.5% | GPT-5.4 mini |

### Reasoning

- Winner: Coming soon
- GPT-5.4 mini public-lane score: 73.9 (Unranked · 2 rankable rows)
- GPT-5.4 nano public-lane score: 73.7 (Unranked · 2 rankable rows)

### Knowledge

- Winner: GPT-5.4 mini
- GPT-5.4 mini public-lane score: 55.6 (Supported · #52/183)
- GPT-5.4 nano public-lane score: 47.4 (Supported · #97/183)

| Benchmark | GPT-5.4 mini | GPT-5.4 nano | Winner |
|-----------|-----------|-----------|--------|
| GPQA | 88% | 82.8% | GPT-5.4 mini |
| HLE | 41.5% | 37.7% | GPT-5.4 mini |
| HLE w/o tools | 28.2% | 24.3% | GPT-5.4 mini |
| GPQA Diamond (Vals) | 83.1% | 77.5% | GPT-5.4 mini |
| MMLU-Pro (Vals) | 84.6% | 77.2% | GPT-5.4 mini |

### Instruction Following

- Winner: Coming soon
- GPT-5.4 mini public-lane score: 89.8 (#23/123)
- GPT-5.4 nano public-lane score: 93.2 (#9/123)

### Mathematics

- Winner: Coming soon
- GPT-5.4 mini public-lane score: 44.5 (Unranked · 2 rankable rows)
- GPT-5.4 nano public-lane score: 43.9 (Unranked · 2 rankable rows)

| Benchmark | GPT-5.4 mini | GPT-5.4 nano | Winner |
|-----------|-----------|-----------|--------|
| FrontierMath v2 (Tiers 1-3) | 28.280% | 25.860% | GPT-5.4 mini |
| FrontierMath v2 (Tier 4) | 2.080% | 6.250% | GPT-5.4 nano |

## FAQ

### Which is better overall, GPT-5.4 mini or GPT-5.4 nano?

GPT-5.4 mini is ahead overall on BenchLM right now.

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

The widest category gap is in multimodal & grounded, where the averages are 57.2 for GPT-5.4 mini and 23.8 for GPT-5.4 nano.

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

GPT-5.4 mini currently has 19 sourced benchmark scores on BenchLM.

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

GPT-5.4 nano currently has 18 sourced benchmark scores on BenchLM.

## Related Comparisons

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

## Explore More

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