# Grok 4.5 vs o3

> Grok 4.5 is the stronger model overall, scoring 55.0 to 47.4 on the Noometry Index.

- Canonical page: https://noometry.com/compare/grok-4-5-vs-o3
- Last updated: 2026-10-10
- Shared benchmarks: 35

## Summary

- They share 35 benchmarks with published results for both. Grok 4.5 scores higher in 8 categories and o3 in 2 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.5 leads 56.0 to 32.1.
- The biggest single-benchmark swing is ARC-AGI-2: 52.6% for Grok 4.5 and 6.5% for o3.
- Grok 4.5 is cheaper at $2 / $6 per million input/output tokens, against $2 / $8 for o3.
- Grok 4.5 accepts more context: 500K tokens versus 200K.

## Snapshot

| | Grok 4.5 | o3 |
|---|---|---|
| Provider | xAI | OpenAI |
| Noometry Index | 55.0 | 47.4 |
| Rank | 22 | 53 |
| Context | 500K | 200K |
| Input $/M | $2 | $2 |
| Output $/M | $6 | $8 |
| Weights | Proprietary | Proprietary |

## Coding

- Grok 4.5: 51.7 (#34)
- o3: 47.3 (#57)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| WeirdML | 46.4% | 52.4% |
| LMArena Coding | 1473 | 1408 |
| ALE-Bench | 1,309 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| DeepSWE | 53.8% | — |
| FrontierCode | 42.4% | — |
| SWE-bench Verified (bash only) | — | 58.4% |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1552 | — |
| SciCode | 54.1% | — |
| GSO | — | 8.8% |
| CadEval | — | 74% |

## Agentic & Tool Use

- Grok 4.5: 44.4 (#15)
- o3: 34.3 (#42)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| LMArena Search | 1213 | 1144 |
| APEX-Agents | 56.2% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| τ²-bench Banking | 47.9% | — |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| PostTrainBench | 23.4% | — |
| GBAEval | 65.4% | — |
| GDP.pdf | 14% | — |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | 3,887 | — |

## Reasoning

- Grok 4.5: 56.0 (#23)
- o3: 32.1 (#66)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| ARC-AGI-2 | 52.6% | 6.5% |
| SimpleBench | 70% | 53.1% |
| Kagi LLM Benchmark | 83.5% | 67.6% |
| ARC-AGI-1 | 87.2% | 60.8% |
| CritPt | 15.4% | 1.4% |
| Chess Puzzles | 36% | 38% |
| LMArena Hard Prompts | 1462 | 1402 |
| DTBench | 96.5% | 84.8% |
| LMCA | 45.2% | 39.7% |
| Epoch Capabilities Index | 154.02 | 146.88 |
| NYT Connections (extended) | 79.9% | — |
| EnigmaEval | — | 13.1% |
| Mystery Game Puzzles | — | 29% |
| Surface Evolver Bench | 74.4% | — |
| ForecastBench | — | 62.5 |

## Math

- Grok 4.5: 60.7 (#29)
- o3: 48.9 (#49)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 57.2% | 33.3% |
| OTIS Mock AIME 2024-2025 | 97.8% | 84.4% |
| LMArena Math | 1457 | 1426 |
| FrontierMath Tier 4 | 24.4% | — |
| ProofBench | 31% | — |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |

## Knowledge

- Grok 4.5: 62.3 (#21)
- o3: 54.5 (#47)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| GPQA Diamond | 93.4% | 81.8% |
| SimpleQA Verified | 48.3% | 49.4% |
| LMArena Expert | 1466 | 1403 |
| Humanity's Last Exam | — | 20.3% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |

## Multimodal

- Grok 4.5: 37.8 (#63)
- o3: 40.7 (#35)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| LMArena Vision | 1287 | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
| Blueprint-Bench 2 | 27.3% | — |
| Furniture Assembly | 22.5% | — |
| LMArena Document | 1452 | — |

## Multilingual

- Grok 4.5: 54.5 (#43)
- o3: 51.9 (#99)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| LMArena Non-English | 1439 | 1402 |
| LMArena Chinese | 1497 | 1438 |
| LMArena French | 1458 | 1430 |
| LMArena German | 1448 | 1420 |
| LMArena Japanese | 1425 | 1404 |
| LMArena Korean | 1402 | 1371 |
| LMArena Russian | 1447 | 1407 |
| LMArena Spanish | 1452 | 1395 |

## Instruction Following

- Grok 4.5: 76.3 (#45)
- o3: 73.0 (#122)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| LMArena Instruction Following | 1446 | 1368 |
| IFEval | — | 86.9% |

## Long Context

- Grok 4.5: 44.8 (#54)
- o3: 53.2 (#6)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| LMArena Longer Query | 1463 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |

## Writing & Preference

- Grok 4.5: 65.9 (#40)
- o3: 63.7 (#62)

| Benchmark | Grok 4.5 | o3 |
|---|---|---|
| LMArena Text | 1448 | 1410 |
| LMArena Creative Writing | 1442 | 1359 |
| EQ-Bench Creative Writing | 1579 | 1676 |
| LMArena Multi-Turn | 1457 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| WildBench | — | 86.1% |

## FAQ

### Is Grok 4.5 better than o3?

Grok 4.5 is the stronger model overall, scoring 55.0 to 47.4 on the Noometry Index.

### Which is cheaper, Grok 4.5 or o3?

Grok 4.5 is cheaper. It lists at $2 per million input tokens and $6 per million output tokens; o3 lists at $2 and $8.

### Is Grok 4.5 or o3 better for coding?

Grok 4.5 scores higher on coding benchmarks: 51.7 versus 47.3 in the Noometry coding category.

### Which has the bigger context window?

Grok 4.5 does, with 500K tokens against 200K.

### How many benchmarks do Grok 4.5 and o3 share?

35 benchmarks have published results for both models. Grok 4.5 has 52 scored results on Noometry and o3 has 63.