# DeepSeek V4 Pro vs GLM-4.7

> DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 42.0 on the Noometry Index.

- Canonical page: https://noometry.com/compare/deepseek-v4-pro-vs-glm-4-7
- Last updated: 2026-10-10
- Shared benchmarks: 31

## Summary

- They share 31 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 9 categories and GLM-4.7 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 24.3.
- The biggest single-benchmark swing is ProofBench: 50% for DeepSeek V4 Pro and 6% for GLM-4.7.
- Both cost about the same: $0.66 input and $1.98 output per million tokens.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 205K.

## Snapshot

| | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| Provider | DeepSeek | Z.ai (Zhipu) |
| Noometry Index | 54.3 | 42.0 |
| Rank | 31 | 124 |
| Context | 1M | 205K |
| Input $/M | $0.66 | $0.60 |
| Output $/M | $1.98 | $2.20 |
| Weights | Open | Open |

## Coding

- DeepSeek V4 Pro: 52.4 (#34)
- GLM-4.7: 44.0 (#79)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| LMArena WebDev | 1582 | 1435 |
| SciCode | 51% | 45.1% |
| LMArena Coding | 1470 | 1454 |
| ALE-Bench | 1,403 | 399.48 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| WeirdML | 66.2% | — |

## Agentic & Tool Use

- DeepSeek V4 Pro: 32.8 (#58)
- GLM-4.7: 26.5 (#103)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| Vending-Bench 2 | 3,285 | 2,377 |
| Terminal-Bench | — | 33.4% |
| APEX-Agents | 47.3% | — |

## Reasoning

- DeepSeek V4 Pro: 56.5 (#24)
- GLM-4.7: 24.3 (#164)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| CritPt | 18% | 1.7% |
| Chess Puzzles | 47% | 6% |
| LMArena Hard Prompts | 1461 | 1443 |
| Epoch Capabilities Index | 155.31 | 143.51 |
| ARC-AGI-2 | 61.3% | — |
| SimpleBench | — | 47.7% |
| Kagi LLM Benchmark | 53.5% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 90.5% | — |
| Mystery Game Puzzles | 43% | — |
| DTBench | 93.9% | — |
| LMCA | 45.5% | — |
| Surface Evolver Bench | 40% | — |
| ForecastBench | 56.1 | — |

## Math

- DeepSeek V4 Pro: 64.8 (#30)
- GLM-4.7: 38.6 (#135)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 98.6% | 83.3% |
| ProofBench | 50% | 6% |
| LMArena Math | 1455 | 1423 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| FrontierMath (Feb 2025 set) | — | 2.4% |
| FrontierMath Tier 4 (v1) | — | 0% |

## Knowledge

- DeepSeek V4 Pro: 59.5 (#31)
- GLM-4.7: 47.0 (#80)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| GPQA Diamond | 91.7% | 83.3% |
| SimpleQA Verified | 52.9% | 32.2% |
| Vectara Hallucination Rate | 8.6% | 11.7% |
| LMArena Expert | 1464 | 1424 |

## Multilingual

- DeepSeek V4 Pro: 54.4 (#45)
- GLM-4.7: 52.8 (#79)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| LMArena Non-English | 1439 | 1417 |
| LMArena Chinese | 1486 | 1495 |
| LMArena French | 1472 | 1432 |
| LMArena German | 1458 | 1424 |
| LMArena Japanese | 1445 | 1439 |
| LMArena Korean | 1447 | 1399 |
| LMArena Russian | 1453 | 1423 |
| LMArena Spanish | 1458 | 1434 |

## Instruction Following

- DeepSeek V4 Pro: 76.1 (#47)
- GLM-4.7: 74.4 (#95)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| LMArena Instruction Following | 1448 | 1411 |

## Long Context

- DeepSeek V4 Pro: 45.0 (#51)
- GLM-4.7: 42.8 (#116)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| CL-bench Life | 13.5% | 10.9% |
| LMArena Longer Query | 1458 | 1432 |
| CL-bench | — | 15.9% |

## Writing & Preference

- DeepSeek V4 Pro: 65.5 (#46)
- GLM-4.7: 60.9 (#93)

| Benchmark | DeepSeek V4 Pro | GLM-4.7 |
|---|---|---|
| LMArena Text | 1451 | 1435 |
| LMArena Creative Writing | 1446 | 1401 |
| EQ-Bench Creative Writing | 1553 | 1413 |
| LMArena Multi-Turn | 1467 | 1446 |
| EQ-Bench 4 | 1166 | — |

## FAQ

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

DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 42.0 on the Noometry Index.

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

DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; GLM-4.7 lists at $0.60 and $2.20.

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

DeepSeek V4 Pro scores higher on coding benchmarks: 52.4 versus 44.0 in the Noometry coding category.

### Which has the bigger context window?

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

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

31 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and GLM-4.7 has 36.