# GPT-5.4 mini vs Grok 4.20 (Non-Reasoning)

> Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.5 to 44.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-5-4-mini-vs-grok-4-20
- Last updated: 2026-10-10
- Shared benchmarks: 38

## Summary

- They share 38 benchmarks with published results for both. GPT-5.4 mini scores higher in 2 categories and Grok 4.20 (Non-Reasoning) in 8 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.20 (Non-Reasoning) leads 52.2 to 30.3.
- The biggest single-benchmark swing is ARC-AGI-2: 18.9% for GPT-5.4 mini and 65.1% for Grok 4.20 (Non-Reasoning).
- Grok 4.20 (Non-Reasoning) is cheaper at $1.25 / $2.50 per million input/output tokens, against $0.75 / $4.50 for GPT-5.4 mini.
- Grok 4.20 (Non-Reasoning) accepts more context: 1M tokens versus 400K.

## Snapshot

| | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 44.8 | 48.5 |
| Rank | 69 | 48 |
| Context | 400K | 1M |
| Input $/M | $0.75 | $1.25 |
| Output $/M | $4.50 | $2.50 |
| Weights | Proprietary | Proprietary |

## Coding

- GPT-5.4 mini: 45.0 (#67)
- Grok 4.20 (Non-Reasoning): 41.8 (#104)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena WebDev | 1397 | 1374 |
| WeirdML | 60.3% | 52.3% |
| LMArena Coding | 1438 | 1458 |
| ALE-Bench | 1,189 | 1,150 |
| FrontierCode | 27% | — |
| SciCode | 49.9% | — |

## Agentic & Tool Use

- GPT-5.4 mini: 29.6 (#74)
- Grok 4.20 (Non-Reasoning): 34.3 (#43)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| Terminal-Bench | — | 57.3% |
| τ²-bench Banking | — | 18% |
| DeepResearch Bench | 36.3% | — |
| LMArena Search | — | 1189 |
| Vending-Bench 2 | — | 4,663 |

## Reasoning

- GPT-5.4 mini: 30.3 (#77)
- Grok 4.20 (Non-Reasoning): 52.2 (#30)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| ARC-AGI-2 | 18.9% | 65.1% |
| Kagi LLM Benchmark | 37.9% | 75% |
| NYT Connections (extended) | 61.8% | 85.4% |
| ARC-AGI-1 | 63.7% | 89.5% |
| Chess Puzzles | 24% | 24% |
| Thematic Generalization | 61.7% | 63.8% |
| LMArena Hard Prompts | 1424 | 1451 |
| DTBench | 80% | 90.1% |
| LMCA | 40.8% | 38.7% |
| Epoch Capabilities Index | 148.83 | 152.01 |
| ForecastBench | 57 | 61.4 |
| CritPt | 10% | — |
| Mystery Game Puzzles | 11% | — |

## Math

- GPT-5.4 mini: 45.4 (#63)
- Grok 4.20 (Non-Reasoning): 48.0 (#54)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| FrontierMath (Tiers 1-3) | 51.2% | 44.9% |
| FrontierMath Tier 4 | 9.8% | 17.1% |
| OTIS Mock AIME 2024-2025 | 88.9% | 92.2% |
| ProofBench | 21% | 14% |
| LMArena Math | 1418 | 1455 |
| FrontierMath (Feb 2025 set) | 28.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |

## Knowledge

- GPT-5.4 mini: 51.4 (#62)
- Grok 4.20 (Non-Reasoning): 52.8 (#54)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| GPQA Diamond | 86.9% | 89.3% |
| SimpleQA Verified | 29.4% | 30.2% |
| LMArena Expert | 1434 | 1440 |
| Vectara Hallucination Rate | 5.5% | — |

## Multimodal

- GPT-5.4 mini: 39.0 (#51)
- Grok 4.20 (Non-Reasoning): 33.2 (#88)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Vision | 1245 | 1263 |
| Blueprint-Bench 2 | — | 0% |
| LMArena Document | — | 1416 |

## Multilingual

- GPT-5.4 mini: 52.1 (#94)
- Grok 4.20 (Non-Reasoning): 54.7 (#38)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Non-English | 1405 | 1440 |
| LMArena Chinese | 1446 | 1481 |
| LMArena French | 1440 | 1477 |
| LMArena German | 1407 | 1463 |
| LMArena Japanese | 1375 | 1450 |
| LMArena Korean | 1369 | 1417 |
| LMArena Russian | 1417 | 1457 |
| LMArena Spanish | 1403 | 1442 |

## Instruction Following

- GPT-5.4 mini: 74.3 (#99)
- Grok 4.20 (Non-Reasoning): 75.0 (#79)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Instruction Following | 1405 | 1420 |

## Long Context

- GPT-5.4 mini: 42.9 (#107)
- Grok 4.20 (Non-Reasoning): 45.4 (#33)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Longer Query | 1407 | 1436 |
| CL-bench | — | 22.2% |
| CL-bench Life | — | 11.9% |

## Writing & Preference

- GPT-5.4 mini: 64.2 (#56)
- Grok 4.20 (Non-Reasoning): 65.9 (#42)

| Benchmark | GPT-5.4 mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Text | 1411 | 1451 |
| LMArena Creative Writing | 1370 | 1439 |
| EQ-Bench Creative Writing | 1665 | 1574 |
| LMArena Multi-Turn | 1429 | 1456 |

## FAQ

### Is GPT-5.4 mini better than Grok 4.20 (Non-Reasoning)?

Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.5 to 44.8 on the Noometry Index.

### Which is cheaper, GPT-5.4 mini or Grok 4.20 (Non-Reasoning)?

Grok 4.20 (Non-Reasoning) is cheaper. It lists at $1.25 per million input tokens and $2.50 per million output tokens; GPT-5.4 mini lists at $0.75 and $4.50.

### Is GPT-5.4 mini or Grok 4.20 (Non-Reasoning) better for coding?

GPT-5.4 mini scores higher on coding benchmarks: 45.0 versus 41.8 in the Noometry coding category.

### Which has the bigger context window?

Grok 4.20 (Non-Reasoning) does, with 1M tokens against 400K.

### How many benchmarks do GPT-5.4 mini and Grok 4.20 (Non-Reasoning) share?

38 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and Grok 4.20 (Non-Reasoning) has 46.