# GPT-3.5-turbo vs Grok 4

> Grok 4 is the stronger model overall, scoring 48.1 to 23.2 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-3-5-turbo-vs-grok-4
- Last updated: 2026-10-10
- Shared benchmarks: 24

## Summary

- They share 24 benchmarks with published results for both. GPT-3.5-turbo scores higher in 0 categories and Grok 4 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok 4 leads 53.8 to 10.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 2.2% for GPT-3.5-turbo and 84% for Grok 4.

## Snapshot

| | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 23.2 | 48.1 |
| Rank | 350 | 56 |
| Context | 16K | — |
| Input $/M | $0.50 | — |
| Output $/M | $1.50 | — |
| Weights | Proprietary | Proprietary |

## Coding

- GPT-3.5-turbo: 23.9 (#331)
- Grok 4: 50.3 (#46)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| WeirdML | 3.5% | 45.7% |
| LMArena Coding | 1136 | 1408 |
| Aider Polyglot | — | 79.6% |
| BigCodeBench Instruct | 39.1% | — |
| BigCodeBench Complete | 50.6% | — |
| HumanEval+ | 70.7% | — |
| MBPP+ | 69.7% | — |

## Agentic & Tool Use

- GPT-3.5-turbo: —
- Grok 4: 32.3 (#68)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| METR Time Horizons | 21.5% | 66.6% |
| Terminal-Bench | — | 27.2% |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 21.1% |
| Cybench | — | 43% |
| DeepResearch Bench | — | 47.3% |
| BALROG | — | 43.6% |
| LMArena Search | — | 1142 |

## Reasoning

- GPT-3.5-turbo: 13.8 (#332)
- Grok 4: 36.7 (#65)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| Chess Puzzles | 0% | 28% |
| LMArena Hard Prompts | 1108 | 1409 |
| Epoch Capabilities Index | 118.55 | 146.44 |
| ForecastBench | 50.4 | 60.9 |
| ARC-AGI-2 | — | 16% |
| SimpleBench | — | 60.5% |
| Kagi LLM Benchmark | — | 73.6% |
| ARC-AGI-1 | — | 66.7% |
| Mystery Game Puzzles | 3% | — |
| DTBench | 48.5% | — |
| LMCA | 9.7% | — |
| Adversarial NLI | 58.1% | — |
| BIG-Bench Hard | 61.6% | — |
| CommonsenseQA 2.0 | 57% | — |
| WinoGrande | 81.6% | — |

## Math

- GPT-3.5-turbo: 6.3 (#327)
- Grok 4: 48.4 (#64)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.2% | 84% |
| LMArena Math | 1142 | 1422 |
| FrontierMath (Tiers 1-3) | 0% | — |
| Omni-MATH | — | 60.3% |
| MATH Level 5 | 15.9% | — |
| FrontierMath (Feb 2025 set) | — | 19.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
| GSM8K | 57.8% | — |

## Knowledge

- GPT-3.5-turbo: 10.0 (#303)
- Grok 4: 53.8 (#55)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| GPQA Diamond | 28% | 87% |
| LMArena Expert | 1070 | 1415 |
| MMLU-Pro | — | 85.1% |
| Confabulations | — | 12.4% |
| GPQA (HELM) | — | 72.7% |
| ARC (AI2) Challenge | 87.4% | — |
| BoolQ | 87% | — |
| MMLU | 71.4% | — |
| OpenBookQA | 86% | — |
| TriviaQA | 85.8% | — |

## Multimodal

- GPT-3.5-turbo: —
- Grok 4: 33.7 (#94)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| LMArena Vision | — | 1210 |
| GeoBench | — | 45% |

## Multilingual

- GPT-3.5-turbo: 31.5 (#258)
- Grok 4: 51.8 (#103)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| LMArena Non-English | 1108 | 1403 |
| LMArena Chinese | 1075 | 1427 |
| LMArena French | 1118 | 1418 |
| LMArena German | 1090 | 1429 |
| LMArena Japanese | 1043 | 1394 |
| LMArena Korean | 1019 | 1377 |
| LMArena Russian | 1123 | 1410 |
| LMArena Spanish | 1121 | 1420 |

## Instruction Following

- GPT-3.5-turbo: 57.9 (#262)
- Grok 4: 79.2 (#5)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| LMArena Instruction Following | 1119 | 1387 |
| IFEval | — | 94.9% |

## Long Context

- GPT-3.5-turbo: 34.0 (#254)
- Grok 4: 63.1 (#4)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| LMArena Longer Query | 1121 | 1409 |
| Fiction.LiveBench | — | 94.4% |

## Writing & Preference

- GPT-3.5-turbo: 25.3 (#305)
- Grok 4: 58.5 (#116)

| Benchmark | GPT-3.5-turbo | Grok 4 |
|---|---|---|
| LMArena Text | 1125 | 1411 |
| LMArena Creative Writing | 1092 | 1397 |
| LMArena Multi-Turn | 1117 | 1416 |
| Short-Story Creative Writing | — | 76.9% |
| EQ-Bench Creative Writing | 451 | — |
| WildBench | — | 79.7% |

## FAQ

### Is GPT-3.5-turbo better than Grok 4?

Grok 4 is the stronger model overall, scoring 48.1 to 23.2 on the Noometry Index.

### Is GPT-3.5-turbo or Grok 4 better for coding?

Grok 4 scores higher on coding benchmarks: 50.3 versus 23.9 in the Noometry coding category.

### How many benchmarks do GPT-3.5-turbo and Grok 4 share?

24 benchmarks have published results for both models. GPT-3.5-turbo has 44 scored results on Noometry and Grok 4 has 48.