# GLM-4.5-Air vs GPT-4.5

> GLM-4.5-Air is the stronger model overall, scoring 38.9 to 37.2 on the Noometry Index.

- Canonical page: https://noometry.com/compare/glm-4-5-air-vs-gpt-4-5
- Last updated: 2026-10-12
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. GLM-4.5-Air scores higher in 4 categories and GPT-4.5 in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GLM-4.5-Air leads 24.1 to 13.9.
- GLM-4.5-Air has downloadable open weights; the other is API-only.

## Snapshot

| | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 38.9 | 37.2 |
| Rank | 177 | 208 |
| Context | 131K | — |
| Input $/M | $0.20 | — |
| Output $/M | $1.10 | — |
| Weights | Open | Proprietary |

## Coding

- GLM-4.5-Air: 33.3 (#259)
- GPT-4.5: 42.2 (#109)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| LMArena Coding | 1397 | 1396 |
| Aider Polyglot | — | 44.9% |
| GSO | 2.9% | — |
| WeirdML | — | 39.4% |
| LiveBench Coding | — | 75.2% |

## Agentic & Tool Use

- GLM-4.5-Air: —
- GPT-4.5: 27.9 (#97)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| Cybench | — | 17.5% |

## Reasoning

- GLM-4.5-Air: 24.1 (#166)
- GPT-4.5: 13.9 (#330)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| LMArena Hard Prompts | 1379 | 1403 |
| ForecastBench | 59.2 | 61.7 |
| ARC-AGI-2 | — | 0.8% |
| SimpleBench | — | 34.5% |
| Kagi LLM Benchmark | 43% | — |
| ARC-AGI-1 | — | 10.3% |
| EnigmaEval | — | 3.2% |
| LiveBench Reasoning | — | 71.1% |
| LiveBench Data Analysis | — | 64.3% |
| Epoch Capabilities Index | — | 136.74 |
| LiveBench | — | 69% |

## Math

- GLM-4.5-Air: 36.2 (#170)
- GPT-4.5: 32.6 (#211)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| LMArena Math | 1396 | 1412 |
| OTIS Mock AIME 2024-2025 | — | 37.8% |
| Omni-MATH | 39.1% | — |
| LiveBench Math | — | 69.3% |
| MATH Level 5 | — | 78.6% |

## Knowledge

- GLM-4.5-Air: 35.0 (#191)
- GPT-4.5: 32.5 (#211)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| Humanity's Last Exam | 8.1% | 5.4% |
| LMArena Expert | 1370 | 1394 |
| GPQA Diamond | — | 68.7% |
| MMLU-Pro | 76.2% | — |
| Confabulations | — | 13.6% |
| Vectara Hallucination Rate | 9.3% | — |
| GPQA (HELM) | 59.4% | — |

## Multimodal

- GLM-4.5-Air: —
- GPT-4.5: 37.6 (#71)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| LMArena Vision | — | 1195 |
| VPCT | — | 45% |

## Multilingual

- GLM-4.5-Air: 49.1 (#135)
- GPT-4.5: 52.5 (#83)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| LMArena Non-English | 1366 | 1413 |
| LMArena Chinese | 1426 | 1421 |
| LMArena French | 1399 | 1418 |
| LMArena German | 1377 | 1457 |
| LMArena Japanese | 1348 | 1416 |
| LMArena Korean | 1308 | 1392 |
| LMArena Russian | 1373 | 1419 |
| LMArena Spanish | 1386 | — |

## Instruction Following

- GLM-4.5-Air: 69.6 (#171)
- GPT-4.5: 72.6 (#134)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| LMArena Instruction Following | 1354 | 1404 |
| LiveBench Instruction Following | — | 72.3% |
| IFEval | 81.2% | — |

## Long Context

- GLM-4.5-Air: 41.6 (#135)
- GPT-4.5: 40.4 (#155)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| LMArena Longer Query | 1366 | 1406 |
| Fiction.LiveBench | — | 63.9% |

## Writing & Preference

- GLM-4.5-Air: 55.9 (#139)
- GPT-4.5: 56.9 (#134)

| Benchmark | GLM-4.5-Air | GPT-4.5 |
|---|---|---|
| LMArena Text | 1384 | 1417 |
| LMArena Creative Writing | 1343 | 1394 |
| LMArena Multi-Turn | 1371 | 1444 |
| Short-Story Creative Writing | — | 75.6% |
| EQ-Bench Creative Writing | — | 1258 |
| WildBench | 78.9% | — |
| LiveBench Language | — | 61.5% |

## FAQ

### Is GLM-4.5-Air better than GPT-4.5?

GLM-4.5-Air is the stronger model overall, scoring 38.9 to 37.2 on the Noometry Index.

### Is GLM-4.5-Air or GPT-4.5 better for coding?

GPT-4.5 scores higher on coding benchmarks: 42.2 versus 33.3 in the Noometry coding category.

### How many benchmarks do GLM-4.5-Air and GPT-4.5 share?

18 benchmarks have published results for both models. GLM-4.5-Air has 27 scored results on Noometry and GPT-4.5 has 42.