# GPT-5 Mini vs Grok 4.20 (Non-Reasoning)

> Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 41.8 on the Noometry Index. GPT-5 Mini costs 2.3× less per token, which makes it the better buy when Grok 4.20 (Non-Reasoning)'s lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gpt-5-mini-vs-grok-4-20
- Last updated: 2026-10-10
- Shared benchmarks: 37

## Summary

- They share 37 benchmarks with published results for both. GPT-5 Mini scores higher in 2 categories and Grok 4.20 (Non-Reasoning) in 8 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.20 (Non-Reasoning) leads 52.3 to 23.9.
- The biggest single-benchmark swing is ARC-AGI-2: 4.4% for GPT-5 Mini and 65.1% for Grok 4.20 (Non-Reasoning).
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $1.25 / $2.50 for Grok 4.20 (Non-Reasoning).
- Grok 4.20 (Non-Reasoning) accepts more context: 1M tokens versus 400K.

## Snapshot

| | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 41.8 | 48.6 |
| Rank | 128 | 54 |
| Context | 400K | 1M |
| Input $/M | $0.25 | $1.25 |
| Output $/M | $2 | $2.50 |
| Weights | Proprietary | Proprietary |

## Coding

- GPT-5 Mini: 40.1 (#146)
- Grok 4.20 (Non-Reasoning): 42.1 (#112)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| WeirdML | 52.7% | 52.3% |
| LMArena Coding | 1406 | 1459 |
| ALE-Bench | 799.77 | 1,150 |
| SWE-bench Verified | 64.7% | — |
| SWE-bench Verified (bash only) | 59.8% | — |
| LMArena WebDev | — | 1375 |
| SWE-bench Multilingual | 39.7% | — |
| SciCode | 39.2% | — |
| AlgoTune | 1.38 | — |

## Agentic & Tool Use

- GPT-5 Mini: 31.1 (#70)
- Grok 4.20 (Non-Reasoning): 34.4 (#46)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| Terminal-Bench | 34.8% | 57.3% |
| Vending-Bench 2 | -31.18 | 4,663 |
| Berkeley Function Calling Leaderboard | 55.5% | — |
| τ²-bench Banking | — | 18% |
| LMArena Search | — | 1189 |

## Reasoning

- GPT-5 Mini: 23.9 (#168)
- Grok 4.20 (Non-Reasoning): 52.3 (#32)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| ARC-AGI-2 | 4.4% | 65.1% |
| Kagi LLM Benchmark | 70.3% | 75% |
| ARC-AGI-1 | 54.3% | 89.5% |
| Chess Puzzles | 30% | 24% |
| LMArena Hard Prompts | 1380 | 1451 |
| DTBench | 80.5% | 90.1% |
| LMCA | 34.2% | 38.7% |
| Epoch Capabilities Index | 145.52 | 151.98 |
| ForecastBench | 61 | 61.4 |
| NYT Connections (extended) | — | 85.4% |
| CritPt | 0% | — |
| EnigmaEval | 8.2% | — |
| Thematic Generalization | — | 63.8% |
| Mystery Game Puzzles | 10% | — |

## Math

- GPT-5 Mini: 46.7 (#69)
- Grok 4.20 (Non-Reasoning): 48.2 (#65)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| FrontierMath (Tiers 1-3) | 46.7% | 44.9% |
| FrontierMath Tier 4 | 12.2% | 17.1% |
| OTIS Mock AIME 2024-2025 | 86.7% | 92.2% |
| ProofBench | 9% | 14% |
| LMArena Math | 1378 | 1455 |
| Omni-MATH | 72.2% | — |
| MATH Level 5 | 97.8% | — |
| FrontierMath (Feb 2025 set) | 27.2% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |

## Knowledge

- GPT-5 Mini: 45.6 (#86)
- Grok 4.20 (Non-Reasoning): 52.8 (#60)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| GPQA Diamond | 75% | 89.3% |
| SimpleQA Verified | 21.6% | 30.2% |
| LMArena Expert | 1379 | 1439 |
| Humanity's Last Exam | 19.4% | — |
| MMLU-Pro | 83.5% | — |
| Confabulations | 13.3% | — |
| Vectara Hallucination Rate | 12.9% | — |
| GPQA (HELM) | 75.6% | — |

## Multimodal

- GPT-5 Mini: 35.6 (#85)
- Grok 4.20 (Non-Reasoning): 33.3 (#98)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Vision | 1202 | 1263 |
| VPCT | 40.2% | — |
| Blueprint-Bench 2 | — | 0% |
| LMArena Document | — | 1416 |

## Multilingual

- GPT-5 Mini: 48.9 (#137)
- Grok 4.20 (Non-Reasoning): 54.5 (#40)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Non-English | 1363 | 1441 |
| LMArena Chinese | 1385 | 1481 |
| LMArena French | 1386 | 1476 |
| LMArena German | 1366 | 1465 |
| LMArena Japanese | 1341 | 1449 |
| LMArena Korean | 1308 | 1417 |
| LMArena Russian | 1362 | 1458 |
| LMArena Spanish | 1355 | 1443 |

## Instruction Following

- GPT-5 Mini: 76.2 (#46)
- Grok 4.20 (Non-Reasoning): 74.8 (#83)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Instruction Following | 1357 | 1420 |
| IFEval | 92.7% | — |

## Long Context

- GPT-5 Mini: 41.9 (#132)
- Grok 4.20 (Non-Reasoning): 45.5 (#34)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Longer Query | 1355 | 1437 |
| Fiction.LiveBench | 69.4% | — |
| CL-bench | — | 22.2% |
| CL-bench Life | — | 11.9% |

## Writing & Preference

- GPT-5 Mini: 55.2 (#148)
- Grok 4.20 (Non-Reasoning): 65.7 (#44)

| Benchmark | GPT-5 Mini | Grok 4.20 (Non-Reasoning) |
|---|---|---|
| LMArena Text | 1373 | 1451 |
| LMArena Creative Writing | 1325 | 1438 |
| EQ-Bench Creative Writing | 1313 | 1574 |
| LMArena Multi-Turn | 1363 | 1456 |
| Short-Story Creative Writing | 83.1% | — |
| WildBench | 85.5% | — |

## FAQ

### Is GPT-5 Mini better than Grok 4.20 (Non-Reasoning)?

Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 41.8 on the Noometry Index. GPT-5 Mini costs 2.3× less per token, which makes it the better buy when Grok 4.20 (Non-Reasoning)'s lead doesn't matter for your workload.

### Which is cheaper, GPT-5 Mini or Grok 4.20 (Non-Reasoning)?

GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; Grok 4.20 (Non-Reasoning) lists at $1.25 and $2.50.

### Is GPT-5 Mini or Grok 4.20 (Non-Reasoning) better for coding?

Grok 4.20 (Non-Reasoning) scores higher on coding benchmarks: 42.1 versus 40.1 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 Mini and Grok 4.20 (Non-Reasoning) share?

37 benchmarks have published results for both models. GPT-5 Mini has 60 scored results on Noometry and Grok 4.20 (Non-Reasoning) has 46.