# GLM-4.5 vs Grok 3

> GLM-4.5 is the stronger model overall, scoring 42.0 to 39.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/glm-4-5-vs-grok-3
- Last updated: 2026-10-10
- Shared benchmarks: 23

## Summary

- They share 23 benchmarks with published results for both. GLM-4.5 scores higher in 4 categories and Grok 3 in 4 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GLM-4.5 leads 28.6 to 13.7.
- GLM-4.5 has downloadable open weights; the other is API-only.

## Snapshot

| | GLM-4.5 | Grok 3 |
|---|---|---|
| Provider | Z.ai (Zhipu) | xAI |
| Noometry Index | 42.0 | 39.9 |
| Rank | 122 | 157 |
| Context | 131K | — |
| Input $/M | $0.60 | — |
| Output $/M | $2.20 | — |
| Weights | Open | Proprietary |

## Coding

- GLM-4.5: 41.4 (#125)
- Grok 3: 41.9 (#115)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| WeirdML | 40.6% | 37.2% |
| LMArena Coding | 1434 | 1432 |
| SWE-bench Verified (bash only) | 54.2% | — |
| Aider Polyglot | — | 53.3% |
| ALE-Bench | 344.82 | — |
| AlgoTune | 1.52 | — |

## Agentic & Tool Use

- GLM-4.5: —
- Grok 3: 30.5 (#76)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| BALROG | — | 29.5% |

## Reasoning

- GLM-4.5: 28.6 (#100)
- Grok 3: 13.7 (#333)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| Kagi LLM Benchmark | 57.9% | 61.3% |
| LMArena Hard Prompts | 1429 | 1434 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 36.1% |
| ARC-AGI-1 | — | 5.5% |
| Epoch Capabilities Index | — | 138.33 |

## Math

- GLM-4.5: 39.0 (#116)
- Grok 3: 38.0 (#145)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| LMArena Math | 1427 | 1391 |
| OTIS Mock AIME 2024-2025 | — | 55.6% |
| Omni-MATH | — | 46.4% |
| MATH Level 5 | — | 88.7% |
| FrontierMath (Feb 2025 set) | — | 3.8% |
| FrontierMath Tier 4 (v1) | — | 0% |

## Knowledge

- GLM-4.5: 35.9 (#179)
- Grok 3: 46.2 (#82)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| Confabulations | 11.3% | 14.2% |
| LMArena Expert | 1433 | 1421 |
| GPQA Diamond | — | 75.8% |
| Humanity's Last Exam | 8.3% | — |
| MMLU-Pro | — | 78.8% |
| Vectara Hallucination Rate | — | 5.8% |
| GPQA (HELM) | — | 65% |

## Multilingual

- GLM-4.5: 52.8 (#77)
- Grok 3: 52.3 (#87)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| LMArena Non-English | 1417 | 1410 |
| LMArena Chinese | 1465 | 1448 |
| LMArena French | 1418 | 1460 |
| LMArena German | 1407 | 1431 |
| LMArena Japanese | 1415 | 1387 |
| LMArena Korean | 1380 | 1373 |
| LMArena Russian | 1414 | 1416 |
| LMArena Spanish | 1454 | 1417 |

## Instruction Following

- GLM-4.5: 74.1 (#104)
- Grok 3: 75.0 (#73)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| LMArena Instruction Following | 1404 | 1409 |
| IFEval | — | 88.4% |

## Long Context

- GLM-4.5: 38.2 (#201)
- Grok 3: 38.7 (#192)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| Fiction.LiveBench | 58.3% | 58.3% |
| LMArena Longer Query | 1412 | 1439 |

## Writing & Preference

- GLM-4.5: 57.5 (#127)
- Grok 3: 55.8 (#141)

| Benchmark | GLM-4.5 | Grok 3 |
|---|---|---|
| LMArena Text | 1430 | 1426 |
| LMArena Creative Writing | 1395 | 1414 |
| Short-Story Creative Writing | 73.4% | 76.4% |
| EQ-Bench Creative Writing | 1343 | 1186 |
| LMArena Multi-Turn | 1415 | 1425 |
| WildBench | — | 84.9% |

## FAQ

### Is GLM-4.5 better than Grok 3?

GLM-4.5 is the stronger model overall, scoring 42.0 to 39.9 on the Noometry Index.

### Is GLM-4.5 or Grok 3 better for coding?

They score almost the same on coding (41.4 vs 41.9); test both on your own repository before choosing.

### How many benchmarks do GLM-4.5 and Grok 3 share?

23 benchmarks have published results for both models. GLM-4.5 has 27 scored results on Noometry and Grok 3 has 40.