# GLM-4.6 vs GPT-4

> GLM-4.6 is the stronger model overall, scoring 40.9 to 28.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/glm-4-6-vs-gpt-4
- Last updated: 2026-10-10
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. GLM-4.6 scores higher in 8 categories and GPT-4 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GLM-4.6 leads 36.4 to 10.1.
- GLM-4.6 is cheaper at $0.60 / $2.20 per million input/output tokens, against $30 / $60 for GPT-4.
- GLM-4.6 accepts more context: 205K tokens versus 8K.
- GLM-4.6 has downloadable open weights; the other is API-only.

## Snapshot

| | GLM-4.6 | GPT-4 |
|---|---|---|
| Provider | Z.ai (Zhipu) | OpenAI |
| Noometry Index | 40.9 | 28.9 |
| Rank | 132 | 308 |
| Context | 205K | 8K |
| Input $/M | $0.60 | $30 |
| Output $/M | $2.20 | $60 |
| Weights | Open | Proprietary |

## Coding

- GLM-4.6: 39.9 (#137)
- GPT-4: 31.2 (#277)

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| LMArena Coding | 1449 | 1254 |
| SWE-bench Verified (bash only) | 55.4% | — |
| LMArena WebDev | 1339 | — |
| SciCode | 38.4% | — |
| WeirdML | — | 12.4% |
| BigCodeBench Instruct | — | 46% |
| BigCodeBench Complete | — | 57.2% |
| ALE-Bench | 340.82 | — |
| HumanEval+ | — | 79.3% |

## Agentic & Tool Use

- GLM-4.6: 32.1 (#60)
- GPT-4: —

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| Terminal-Bench | 24.5% | — |
| Berkeley Function Calling Leaderboard | 72.4% | — |
| METR Time Horizons | — | 36.1% |

## Reasoning

- GLM-4.6: 23.1 (#168)
- GPT-4: 17.3 (#285)

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| LMArena Hard Prompts | 1440 | 1241 |
| Kagi LLM Benchmark | 47.4% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | — | 4% |
| Mystery Game Puzzles | — | 12% |
| DTBench | — | 62.7% |
| LMCA | — | 17.1% |
| BIG-Bench Hard | — | 75.1% |
| Epoch Capabilities Index | — | 125.88 |
| ForecastBench | — | 57.8 |
| HellaSwag | — | 95.3% |
| WinoGrande | — | 87.5% |

## Math

- GLM-4.6: 36.4 (#115)
- GPT-4: 10.1 (#304)

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| LMArena Math | 1431 | 1269 |
| OTIS Mock AIME 2024-2025 | — | 1.1% |
| MATH Level 5 | — | 23% |
| FrontierMath (Feb 2025 set) | 3.8% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
| GSM8K | — | 92% |

## Knowledge

- GLM-4.6: 39.6 (#118)
- GPT-4: 18.4 (#277)

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| LMArena Expert | 1432 | 1211 |
| GPQA Diamond | — | 35.7% |
| Vectara Hallucination Rate | 9.5% | — |
| MMLU | — | 86.4% |
| TriviaQA | — | 84.8% |

## Multilingual

- GLM-4.6: 53.6 (#62)
- GPT-4: 40.8 (#210)

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| LMArena Non-English | 1426 | 1246 |
| LMArena Chinese | 1500 | 1242 |
| LMArena French | 1459 | 1282 |
| LMArena German | 1447 | 1251 |
| LMArena Japanese | 1394 | 1209 |
| LMArena Korean | 1401 | 1184 |
| LMArena Russian | 1418 | 1251 |
| LMArena Spanish | 1436 | 1261 |

## Instruction Following

- GLM-4.6: 74.5 (#93)
- GPT-4: 65.5 (#217)

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| LMArena Instruction Following | 1410 | 1241 |

## Long Context

- GLM-4.6: 43.4 (#91)
- GPT-4: 37.7 (#207)

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| LMArena Longer Query | 1423 | 1244 |

## Writing & Preference

- GLM-4.6: 61.3 (#86)
- GPT-4: 35.0 (#263)

| Benchmark | GLM-4.6 | GPT-4 |
|---|---|---|
| LMArena Text | 1440 | 1263 |
| LMArena Creative Writing | 1411 | 1244 |
| EQ-Bench Creative Writing | 1411 | 752 |
| LMArena Multi-Turn | 1427 | 1257 |

## FAQ

### Is GLM-4.6 better than GPT-4?

GLM-4.6 is the stronger model overall, scoring 40.9 to 28.9 on the Noometry Index.

### Which is cheaper, GLM-4.6 or GPT-4?

GLM-4.6 is cheaper. It lists at $0.60 per million input tokens and $2.20 per million output tokens; GPT-4 lists at $30 and $60.

### Is GLM-4.6 or GPT-4 better for coding?

GLM-4.6 scores higher on coding benchmarks: 39.9 versus 31.2 in the Noometry coding category.

### Which has the bigger context window?

GLM-4.6 does, with 205K tokens against 8K.

### How many benchmarks do GLM-4.6 and GPT-4 share?

18 benchmarks have published results for both models. GLM-4.6 has 29 scored results on Noometry and GPT-4 has 38.