# DeepSeek V4 Flash vs GLM-4.7

> DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 41.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-flash-vs-glm-4-7
- Last updated: 2026-10-10
- Shared benchmarks: 29

## Summary

- They share 29 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 8 categories and GLM-4.7 in 0 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 23.7.
- The biggest single-benchmark swing is ProofBench: 56% for DeepSeek V4 Flash and 6% for GLM-4.7.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.60 / $2.20 for GLM-4.7.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 205K.

## Snapshot

| | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| Provider | DeepSeek | Z.ai (Zhipu) |
| Noometry Index | 53.6 | 41.7 |
| Rank | 32 | 117 |
| Context | 1M | 205K |
| Input $/M | $0.15 | $0.60 |
| Output $/M | $0.60 | $2.20 |
| Weights | Open | Open |

## Coding

- DeepSeek V4 Flash: 47.6 (#56)
- GLM-4.7: 43.6 (#74)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| LMArena WebDev | 1581 | 1435 |
| SciCode | 49.9% | 45.1% |
| LMArena Coding | 1457 | 1454 |
| ALE-Bench | 1,306 | 399.48 |
| FrontierCode | 18.8% | — |
| WeirdML | 63% | — |

## Agentic & Tool Use

- DeepSeek V4 Flash: —
- GLM-4.7: 26.4 (#96)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| Terminal-Bench | — | 33.4% |
| Vending-Bench 2 | — | 2,377 |

## Reasoning

- DeepSeek V4 Flash: 53.7 (#25)
- GLM-4.7: 23.7 (#161)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| SimpleBench | 61.1% | 47.7% |
| CritPt | 16.6% | 1.7% |
| Chess Puzzles | 33% | 6% |
| LMArena Hard Prompts | 1443 | 1443 |
| Epoch Capabilities Index | 154.51 | 143.53 |
| ARC-AGI-2 | 61.4% | — |
| Kagi LLM Benchmark | 52.2% | — |
| NYT Connections (extended) | 89.6% | — |
| ARC-AGI-1 | 89% | — |
| Mystery Game Puzzles | 34% | — |
| DTBench | 90.9% | — |
| LMCA | 41.7% | — |

## Math

- DeepSeek V4 Flash: 60.0 (#31)
- GLM-4.7: 37.7 (#92)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 94.4% | 83.3% |
| ProofBench | 56% | 6% |
| LMArena Math | 1427 | 1423 |
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| FrontierMath (Feb 2025 set) | — | 2.4% |
| FrontierMath Tier 4 (v1) | — | 0% |

## Knowledge

- DeepSeek V4 Flash: 55.5 (#43)
- GLM-4.7: 47.0 (#74)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| GPQA Diamond | 91% | 83.3% |
| SimpleQA Verified | 33.6% | 32.2% |
| LMArena Expert | 1441 | 1424 |
| Vectara Hallucination Rate | — | 11.7% |

## Multilingual

- DeepSeek V4 Flash: 53.2 (#71)
- GLM-4.7: 53.0 (#75)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| LMArena Non-English | 1420 | 1417 |
| LMArena Chinese | 1467 | 1495 |
| LMArena French | 1438 | 1432 |
| LMArena German | 1418 | 1426 |
| LMArena Japanese | 1403 | 1439 |
| LMArena Korean | 1383 | 1401 |
| LMArena Russian | 1428 | 1423 |
| LMArena Spanish | 1437 | 1433 |

## Instruction Following

- DeepSeek V4 Flash: 75.1 (#76)
- GLM-4.7: 74.6 (#90)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| LMArena Instruction Following | 1421 | 1411 |

## Long Context

- DeepSeek V4 Flash: 43.8 (#81)
- GLM-4.7: 42.8 (#111)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| LMArena Longer Query | 1435 | 1432 |
| CL-bench | — | 15.9% |
| CL-bench Life | — | 10.9% |

## Writing & Preference

- DeepSeek V4 Flash: 64.0 (#59)
- GLM-4.7: 61.1 (#89)

| Benchmark | DeepSeek V4 Flash | GLM-4.7 |
|---|---|---|
| LMArena Text | 1432 | 1436 |
| LMArena Creative Writing | 1403 | 1401 |
| EQ-Bench Creative Writing | 1559 | 1413 |
| LMArena Multi-Turn | 1449 | 1446 |

## FAQ

### Is DeepSeek V4 Flash better than GLM-4.7?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 41.7 on the Noometry Index.

### Which is cheaper, DeepSeek V4 Flash or GLM-4.7?

DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; GLM-4.7 lists at $0.60 and $2.20.

### Is DeepSeek V4 Flash or GLM-4.7 better for coding?

DeepSeek V4 Flash scores higher on coding benchmarks: 47.6 versus 43.6 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek V4 Flash does, with 1M tokens against 205K.

### How many benchmarks do DeepSeek V4 Flash and GLM-4.7 share?

29 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and GLM-4.7 has 36.