# DeepSeek V4 Flash vs GLM-4.5-Air

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 38.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-flash-vs-glm-4-5-air
- Last updated: 2026-10-12
- Shared benchmarks: 18

## Summary

- They share 18 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 8 categories and GLM-4.5-Air in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 24.1.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 52.2% for DeepSeek V4 Flash and 43% for GLM-4.5-Air.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.20 / $1.10 for GLM-4.5-Air.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 131K.

## Snapshot

| | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| Provider | DeepSeek | Z.ai (Zhipu) |
| Noometry Index | 53.6 | 38.9 |
| Rank | 35 | 177 |
| Context | 1M | 131K |
| Input $/M | $0.15 | $0.20 |
| Output $/M | $0.60 | $1.10 |
| Weights | Open | Open |

## Coding

- DeepSeek V4 Flash: 47.9 (#59)
- GLM-4.5-Air: 33.3 (#259)

| Benchmark | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| LMArena Coding | 1457 | 1397 |
| FrontierCode | 18.8% | — |
| LMArena WebDev | 1582 | — |
| SciCode | 49.9% | — |
| GSO | — | 2.9% |
| WeirdML | 63% | — |
| ALE-Bench | 1,306 | — |

## Reasoning

- DeepSeek V4 Flash: 53.7 (#30)
- GLM-4.5-Air: 24.1 (#166)

| Benchmark | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| Kagi LLM Benchmark | 52.2% | 43% |
| LMArena Hard Prompts | 1444 | 1379 |
| ARC-AGI-2 | 61.4% | — |
| SimpleBench | 61.1% | — |
| NYT Connections (extended) | 89.6% | — |
| ARC-AGI-1 | 89% | — |
| CritPt | 16.6% | — |
| Chess Puzzles | 33% | — |
| Mystery Game Puzzles | 34% | — |
| DTBench | 90.9% | — |
| LMCA | 41.7% | — |
| Epoch Capabilities Index | 154.49 | — |
| ForecastBench | — | 59.2 |

## Math

- DeepSeek V4 Flash: 60.3 (#37)
- GLM-4.5-Air: 36.2 (#170)

| Benchmark | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| LMArena Math | 1427 | 1396 |
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| OTIS Mock AIME 2024-2025 | 94.4% | — |
| ProofBench | 56% | — |
| Omni-MATH | — | 39.1% |

## Knowledge

- DeepSeek V4 Flash: 55.4 (#48)
- GLM-4.5-Air: 35.0 (#191)

| Benchmark | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| LMArena Expert | 1441 | 1370 |
| GPQA Diamond | 91% | — |
| Humanity's Last Exam | — | 8.1% |
| SimpleQA Verified | 33.6% | — |
| MMLU-Pro | — | 76.2% |
| Vectara Hallucination Rate | — | 9.3% |
| GPQA (HELM) | — | 59.4% |

## Multilingual

- DeepSeek V4 Flash: 53.0 (#72)
- GLM-4.5-Air: 49.1 (#135)

| Benchmark | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| LMArena Non-English | 1420 | 1366 |
| LMArena Chinese | 1468 | 1426 |
| LMArena French | 1439 | 1399 |
| LMArena German | 1418 | 1377 |
| LMArena Japanese | 1406 | 1348 |
| LMArena Korean | 1384 | 1308 |
| LMArena Russian | 1428 | 1373 |
| LMArena Spanish | 1436 | 1386 |

## Instruction Following

- DeepSeek V4 Flash: 74.9 (#81)
- GLM-4.5-Air: 69.6 (#171)

| Benchmark | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| LMArena Instruction Following | 1421 | 1354 |
| IFEval | — | 81.2% |

## Long Context

- DeepSeek V4 Flash: 43.8 (#85)
- GLM-4.5-Air: 41.6 (#135)

| Benchmark | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| LMArena Longer Query | 1434 | 1366 |

## Writing & Preference

- DeepSeek V4 Flash: 63.8 (#61)
- GLM-4.5-Air: 55.9 (#139)

| Benchmark | DeepSeek V4 Flash | GLM-4.5-Air |
|---|---|---|
| LMArena Text | 1432 | 1384 |
| LMArena Creative Writing | 1403 | 1343 |
| LMArena Multi-Turn | 1449 | 1371 |
| EQ-Bench Creative Writing | 1559 | — |
| WildBench | — | 78.9% |

## FAQ

### Is DeepSeek V4 Flash better than GLM-4.5-Air?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 38.9 on the Noometry Index.

### Which is cheaper, DeepSeek V4 Flash or GLM-4.5-Air?

DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; GLM-4.5-Air lists at $0.20 and $1.10.

### Is DeepSeek V4 Flash or GLM-4.5-Air better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.9 versus 33.3 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Flash does, with 1M tokens against 131K.

### How many benchmarks do DeepSeek V4 Flash and GLM-4.5-Air share?

18 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and GLM-4.5-Air has 27.