# GPT-5.1 vs Grok 4.5

> Grok 4.5 is the stronger model overall, scoring 55.0 to 48.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-5-1-vs-grok-4-5
- Last updated: 2026-10-10
- Shared benchmarks: 36

## Summary

- They share 36 benchmarks with published results for both. GPT-5.1 scores higher in 3 categories and Grok 4.5 in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.5 leads 56.0 to 39.9.
- The biggest single-benchmark swing is ARC-AGI-2: 17.6% for GPT-5.1 and 52.6% for Grok 4.5.
- Grok 4.5 is cheaper at $2 / $6 per million input/output tokens, against $1.25 / $10 for GPT-5.1.
- Grok 4.5 accepts more context: 500K tokens versus 400K.

## Snapshot

| | GPT-5.1 | Grok 4.5 |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 48.8 | 55.0 |
| Rank | 47 | 22 |
| Context | 400K | 500K |
| Input $/M | $1.25 | $2 |
| Output $/M | $10 | $6 |
| Weights | Proprietary | Proprietary |

## Coding

- GPT-5.1: 46.9 (#59)
- Grok 4.5: 51.7 (#34)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| LMArena WebDev | 1395 | 1552 |
| SciCode | 43.3% | 54.1% |
| WeirdML | 60.8% | 46.4% |
| LMArena Coding | 1454 | 1473 |
| ALE-Bench | 1,192 | 1,309 |
| SWE-bench Verified | 68% | — |
| DeepSWE | — | 53.8% |
| FrontierCode | — | 42.4% |
| SWE-bench Verified (bash only) | 66% | — |
| GSO | 13.7% | — |
| LiveBench Coding | 72.5% | — |

## Agentic & Tool Use

- GPT-5.1: 32.5 (#55)
- Grok 4.5: 44.4 (#15)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| LMArena Search | 1199 | 1213 |
| Vending-Bench 2 | 1,473 | 3,887 |
| Terminal-Bench | 47.6% | — |
| APEX-Agents | — | 56.2% |
| τ²-bench Banking | — | 47.9% |
| DeepResearch Bench | 42.8% | — |
| PostTrainBench | — | 23.4% |
| GBAEval | — | 65.4% |
| GDP.pdf | — | 14% |

## Reasoning

- GPT-5.1: 39.9 (#50)
- Grok 4.5: 56.0 (#23)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| ARC-AGI-2 | 17.6% | 52.6% |
| SimpleBench | 53.2% | 70% |
| ARC-AGI-1 | 72.8% | 87.2% |
| CritPt | 4.9% | 15.4% |
| Chess Puzzles | 32% | 36% |
| LMArena Hard Prompts | 1457 | 1462 |
| DTBench | 90.1% | 96.5% |
| LMCA | 43.9% | 45.2% |
| Epoch Capabilities Index | 149.66 | 154.02 |
| Kagi LLM Benchmark | — | 83.5% |
| NYT Connections (extended) | — | 79.9% |
| EnigmaEval | 11.2% | — |
| LiveBench Reasoning | 95.8% | — |
| Mystery Game Puzzles | 19% | — |
| LiveBench Data Analysis | 72.1% | — |
| Surface Evolver Bench | — | 74.4% |
| ForecastBench | 58.1 | — |
| LiveBench | 78.8% | — |

## Math

- GPT-5.1: 50.3 (#45)
- Grok 4.5: 60.7 (#29)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.6% | 97.8% |
| LMArena Math | 1446 | 1457 |
| FrontierMath (Tiers 1-3) | — | 57.2% |
| FrontierMath Tier 4 | — | 24.4% |
| ProofBench | — | 31% |
| Omni-MATH | 46.4% | — |
| LiveBench Math | 94.5% | — |
| FrontierMath (Feb 2025 set) | 31% | — |
| FrontierMath Tier 4 (v1) | 12.5% | — |

## Knowledge

- GPT-5.1: 50.5 (#65)
- Grok 4.5: 62.3 (#21)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| GPQA Diamond | 87.6% | 93.4% |
| SimpleQA Verified | 48% | 48.3% |
| LMArena Expert | 1470 | 1466 |
| Humanity's Last Exam | 23.7% | — |
| MMLU-Pro | 57.9% | — |
| Vectara Hallucination Rate | 10.9% | — |
| GPQA (HELM) | 44.2% | — |

## Multimodal

- GPT-5.1: 44.0 (#17)
- Grok 4.5: 37.8 (#63)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| LMArena Vision | 1250 | 1287 |
| LMArena Document | 1403 | 1452 |
| VPCT | 58.7% | — |
| Blueprint-Bench 2 | — | 27.3% |
| Furniture Assembly | — | 22.5% |

## Multilingual

- GPT-5.1: 54.0 (#51)
- Grok 4.5: 54.5 (#43)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| LMArena Non-English | 1431 | 1439 |
| LMArena Chinese | 1495 | 1497 |
| LMArena French | 1451 | 1458 |
| LMArena German | 1439 | 1448 |
| LMArena Japanese | 1453 | 1425 |
| LMArena Korean | 1401 | 1402 |
| LMArena Russian | 1435 | 1447 |
| LMArena Spanish | 1430 | 1452 |

## Instruction Following

- GPT-5.1: 84.0 (#1)
- Grok 4.5: 76.3 (#45)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| LMArena Instruction Following | 1442 | 1446 |
| LiveBench Instruction Following | 93.3% | — |
| IFEval | 93.5% | — |

## Long Context

- GPT-5.1: 47.6 (#14)
- Grok 4.5: 44.8 (#54)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| LMArena Longer Query | 1448 | 1463 |
| CL-bench | 23.7% | — |
| CL-bench Life | 17.3% | — |

## Writing & Preference

- GPT-5.1: 64.7 (#53)
- Grok 4.5: 65.9 (#40)

| Benchmark | GPT-5.1 | Grok 4.5 |
|---|---|---|
| LMArena Text | 1443 | 1448 |
| LMArena Creative Writing | 1427 | 1442 |
| LMArena Multi-Turn | 1449 | 1457 |
| EQ-Bench Creative Writing | — | 1579 |
| WildBench | 86.3% | — |
| LiveBench Language | 80.2% | — |

## FAQ

### Is GPT-5.1 better than Grok 4.5?

Grok 4.5 is the stronger model overall, scoring 55.0 to 48.8 on the Noometry Index.

### Which is cheaper, GPT-5.1 or Grok 4.5?

Grok 4.5 is cheaper. It lists at $2 per million input tokens and $6 per million output tokens; GPT-5.1 lists at $1.25 and $10.

### Is GPT-5.1 or Grok 4.5 better for coding?

Grok 4.5 scores higher on coding benchmarks: 51.7 versus 46.9 in the Noometry coding category.

### Which has the bigger context window?

Grok 4.5 does, with 500K tokens against 400K.

### How many benchmarks do GPT-5.1 and Grok 4.5 share?

36 benchmarks have published results for both models. GPT-5.1 has 63 scored results on Noometry and Grok 4.5 has 52.