# GLM-5.1 vs GLM-5.2

> GLM-5.2 is the stronger model overall, scoring 50.9 to 47.6 on the Noometry Index.

- Canonical page: https://noometry.com/compare/glm-5-1-vs-glm-5-2
- Last updated: 2026-10-10
- Shared benchmarks: 36

## Summary

- They share 36 benchmarks with published results for both. GLM-5.1 scores higher in 0 categories and GLM-5.2 in 9 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GLM-5.2 leads 31.6 to 24.8.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 36.8% for GLM-5.1 and 59.2% for GLM-5.2.
- Both cost about the same: $1.40 input and $4.40 output per million tokens.
- GLM-5.2 accepts more context: 1M tokens versus 200K.

## Snapshot

| | GLM-5.1 | GLM-5.2 |
|---|---|---|
| Provider | Z.ai (Zhipu) | Z.ai (Zhipu) |
| Noometry Index | 47.6 | 50.9 |
| Rank | 51 | 38 |
| Context | 200K | 1M |
| Input $/M | $1.40 | $1.40 |
| Output $/M | $4.40 | $4.40 |
| Weights | Open | Open |

## Coding

- GLM-5.1: 48.5 (#52)
- GLM-5.2: 50.9 (#37)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| SWE-bench Verified | 74.2% | 78.7% |
| LMArena WebDev | 1509 | 1604 |
| SciCode | 43.8% | 50.5% |
| WeirdML | 57.1% | 70.1% |
| LMArena Coding | 1486 | 1484 |
| ALE-Bench | 887.1 | 1,047 |
| DeepSWE | — | 43.8% |
| FrontierCode | — | 24.5% |

## Agentic & Tool Use

- GLM-5.1: 24.8 (#106)
- GLM-5.2: 31.6 (#63)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| GBAEval | 0% | 0% |
| Vending-Bench 2 | 5,634 | 8,314 |
| APEX-Agents | 40.9% | — |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 31.7% |
| ExploitBench | 18.1% | — |

## Reasoning

- GLM-5.1: 38.2 (#57)
- GLM-5.2: 42.1 (#46)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| SimpleBench | 55.1% | 58.8% |
| NYT Connections (extended) | 77.7% | 74.3% |
| CritPt | 4.6% | 20.9% |
| Chess Puzzles | 19% | 21% |
| LMArena Hard Prompts | 1472 | 1480 |
| Epoch Capabilities Index | 149.86 | 151.76 |
| ARC-AGI-2 | — | 22.8% |
| Kagi LLM Benchmark | — | 62.6% |
| ARC-AGI-1 | — | 77% |
| Thematic Generalization | 69.8% | — |
| EBR-Bench | — | 9.5% |
| Mystery Game Puzzles | — | 19% |
| DTBench | — | 93.6% |
| LMCA | — | 45.8% |
| Surface Evolver Bench | — | 55.6% |

## Math

- GLM-5.1: 49.0 (#48)
- GLM-5.2: 55.4 (#35)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 36.8% | 59.2% |
| MathArena Final-Answer Competitions | 67.1% | 67.6% |
| OTIS Mock AIME 2024-2025 | 93.3% | 86.4% |
| ProofBench | 22.2% | 35% |
| LMArena Math | 1472 | 1482 |
| FrontierMath Tier 4 | — | 29.3% |
| FrontierMath (Feb 2025 set) | 33.4% | — |
| FrontierMath Tier 4 (v1) | 12.5% | — |

## Knowledge

- GLM-5.1: 55.0 (#45)
- GLM-5.2: 57.1 (#36)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| GPQA Diamond | 89.9% | 91.9% |
| SimpleQA Verified | 34% | 34.2% |
| LMArena Expert | 1477 | 1484 |

## Multilingual

- GLM-5.1: 55.2 (#34)
- GLM-5.2: 56.0 (#25)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| LMArena Non-English | 1448 | 1459 |
| LMArena Chinese | 1516 | 1518 |
| LMArena French | 1478 | 1476 |
| LMArena German | 1466 | 1470 |
| LMArena Japanese | 1436 | 1451 |
| LMArena Korean | 1423 | 1445 |
| LMArena Russian | 1453 | 1467 |
| LMArena Spanish | 1467 | 1477 |

## Instruction Following

- GLM-5.1: 76.5 (#41)
- GLM-5.2: 77.1 (#32)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| LMArena Instruction Following | 1451 | 1465 |

## Long Context

- GLM-5.1: 44.9 (#51)
- GLM-5.2: 45.3 (#43)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| LMArena Longer Query | 1467 | 1479 |

## Writing & Preference

- GLM-5.1: 67.2 (#29)
- GLM-5.2: 70.7 (#20)

| Benchmark | GLM-5.1 | GLM-5.2 |
|---|---|---|
| LMArena Text | 1462 | 1470 |
| LMArena Creative Writing | 1454 | 1465 |
| EQ-Bench Creative Writing | 1592 | 1757 |
| LMArena Multi-Turn | 1473 | 1468 |
| EQ-Bench 4 | — | 1222 |

## FAQ

### Is GLM-5.1 better than GLM-5.2?

GLM-5.2 is the stronger model overall, scoring 50.9 to 47.6 on the Noometry Index.

### Which is cheaper, GLM-5.1 or GLM-5.2?

GLM-5.2 is cheaper. It lists at $1.40 per million input tokens and $4.40 per million output tokens; GLM-5.1 lists at $1.40 and $4.40.

### Is GLM-5.1 or GLM-5.2 better for coding?

GLM-5.2 scores higher on coding benchmarks: 50.9 versus 48.5 in the Noometry coding category.

### Which has the bigger context window?

GLM-5.2 does, with 1M tokens against 200K.

### How many benchmarks do GLM-5.1 and GLM-5.2 share?

36 benchmarks have published results for both models. GLM-5.1 has 41 scored results on Noometry and GLM-5.2 has 50.