# DeepSeek-V3.1 vs GLM-4.6

> DeepSeek-V3.1 is the stronger model overall, scoring 42.2 to 40.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v3-1-vs-glm-4-6
- Last updated: 2026-10-10
- Shared benchmarks: 20

## Summary

- They share 20 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 2 categories and GLM-4.6 in 6 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in long context, where GLM-4.6 leads 43.4 to 36.3.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 53.2% for DeepSeek-V3.1 and 47.4% for GLM-4.6.
- DeepSeek-V3.1 is cheaper at $0.25 / $0.95 per million input/output tokens, against $0.60 / $2.20 for GLM-4.6.
- GLM-4.6 accepts more context: 205K tokens versus 164K.

## Snapshot

| | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| Provider | DeepSeek | Z.ai (Zhipu) |
| Noometry Index | 42.2 | 40.9 |
| Rank | 106 | 132 |
| Context | 164K | 205K |
| Input $/M | $0.25 | $0.60 |
| Output $/M | $0.95 | $2.20 |
| Weights | Open | Open |

## Coding

- DeepSeek-V3.1: 39.8 (#140)
- GLM-4.6: 39.9 (#137)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| LMArena Coding | 1417 | 1449 |
| SWE-bench Verified (bash only) | — | 55.4% |
| LMArena WebDev | — | 1339 |
| SciCode | — | 38.4% |
| WeirdML | 38.4% | — |
| ALE-Bench | — | 340.82 |

## Agentic & Tool Use

- DeepSeek-V3.1: —
- GLM-4.6: 32.1 (#60)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| Terminal-Bench | — | 24.5% |
| Berkeley Function Calling Leaderboard | — | 72.4% |

## Reasoning

- DeepSeek-V3.1: 27.3 (#109)
- GLM-4.6: 23.1 (#168)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 47.4% |
| LMArena Hard Prompts | 1417 | 1440 |
| SimpleBench | 40% | — |
| CritPt | — | 1.1% |
| DTBench | 82.7% | — |
| LMCA | 24.3% | — |
| Epoch Capabilities Index | 139.92 | — |
| ForecastBench | 58 | — |

## Math

- DeepSeek-V3.1: 36.1 (#125)
- GLM-4.6: 36.4 (#115)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| LMArena Math | 1420 | 1431 |
| FrontierMath (Feb 2025 set) | — | 3.8% |
| FrontierMath Tier 4 (v1) | — | 2.1% |

## Knowledge

- DeepSeek-V3.1: 43.0 (#86)
- GLM-4.6: 39.6 (#118)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| Vectara Hallucination Rate | 5.5% | 9.5% |
| LMArena Expert | 1403 | 1432 |

## Multilingual

- DeepSeek-V3.1: 51.8 (#100)
- GLM-4.6: 53.6 (#62)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| LMArena Non-English | 1400 | 1426 |
| LMArena Chinese | 1470 | 1500 |
| LMArena French | 1447 | 1459 |
| LMArena German | 1412 | 1447 |
| LMArena Japanese | 1374 | 1394 |
| LMArena Korean | 1337 | 1401 |
| LMArena Russian | 1405 | 1418 |
| LMArena Spanish | 1429 | 1436 |

## Instruction Following

- DeepSeek-V3.1: 74.1 (#105)
- GLM-4.6: 74.5 (#93)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| LMArena Instruction Following | 1400 | 1410 |

## Long Context

- DeepSeek-V3.1: 36.3 (#227)
- GLM-4.6: 43.4 (#91)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| LMArena Longer Query | 1422 | 1423 |
| Fiction.LiveBench | 52.8% | — |

## Writing & Preference

- DeepSeek-V3.1: 60.5 (#93)
- GLM-4.6: 61.3 (#86)

| Benchmark | DeepSeek-V3.1 | GLM-4.6 |
|---|---|---|
| LMArena Text | 1420 | 1440 |
| LMArena Creative Writing | 1401 | 1411 |
| EQ-Bench Creative Writing | 1436 | 1411 |
| LMArena Multi-Turn | 1408 | 1427 |

## FAQ

### Is DeepSeek-V3.1 better than GLM-4.6?

DeepSeek-V3.1 is the stronger model overall, scoring 42.2 to 40.9 on the Noometry Index.

### Which is cheaper, DeepSeek-V3.1 or GLM-4.6?

DeepSeek-V3.1 is cheaper. It lists at $0.25 per million input tokens and $0.95 per million output tokens; GLM-4.6 lists at $0.60 and $2.20.

### Is DeepSeek-V3.1 or GLM-4.6 better for coding?

They score almost the same on coding (39.8 vs 39.9); test both on your own repository before choosing.

### Which has the bigger context window?

GLM-4.6 does, with 205K tokens against 164K.

### How many benchmarks do DeepSeek-V3.1 and GLM-4.6 share?

20 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and GLM-4.6 has 29.