# GLM-5.2 vs Qwen3 235B-A22B

> GLM-5.2 is the stronger model overall, scoring 51.1 to 43.5 on the Noometry Index. Qwen3 235B-A22B costs 1.8× less per token, which makes it the better buy when GLM-5.2's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/glm-5-2-vs-qwen3-235b-a22b
- Last updated: 2026-10-12
- Shared benchmarks: 34

## Summary

- They share 34 benchmarks with published results for both. GLM-5.2 scores higher in 7 categories and Qwen3 235B-A22B in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GLM-5.2 leads 42.3 to 15.7.
- The biggest single-benchmark swing is ARC-AGI-1: 77% for GLM-5.2 and 11% for Qwen3 235B-A22B.
- Qwen3 235B-A22B is cheaper at $0.70 / $2.80 per million input/output tokens, against $1.40 / $4.40 for GLM-5.2.
- GLM-5.2 accepts more context: 1M tokens versus 131K.

## Snapshot

| | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| Provider | Z.ai (Zhipu) | Alibaba (Qwen) |
| Noometry Index | 51.1 | 43.5 |
| Rank | 44 | 91 |
| Context | 1M | 131K |
| Input $/M | $1.40 | $0.70 |
| Output $/M | $4.40 | $2.80 |
| Weights | Open | Open |

## Coding

- GLM-5.2: 51.3 (#41)
- Qwen3 235B-A22B: 44.3 (#75)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| SciCode | 50.5% | 42.4% |
| WeirdML | 70.1% | 41% |
| LMArena Coding | 1485 | 1445 |
| SWE-bench Verified | 78.7% | — |
| DeepSWE | 43.8% | — |
| FrontierCode | 24.5% | — |
| Aider Polyglot | — | 59.6% |
| LMArena WebDev | 1603 | — |
| ALE-Bench | 1,047 | — |

## Agentic & Tool Use

- GLM-5.2: 32.4 (#63)
- Qwen3 235B-A22B: 33.9 (#51)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| Vending-Bench 2 | 8,314 | -11.34 |
| APEX-Agents | 45.2% | — |
| Berkeley Function Calling Leaderboard | — | 52.1% |
| τ²-bench Banking | 37.1% | — |
| PostTrainBench | 31.7% | — |
| GBAEval | 0% | — |

## Reasoning

- GLM-5.2: 42.3 (#52)
- Qwen3 235B-A22B: 15.7 (#311)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| ARC-AGI-2 | 22.8% | 1.3% |
| SimpleBench | 58.8% | 31% |
| Kagi LLM Benchmark | 62.6% | 69.4% |
| ARC-AGI-1 | 77% | 11% |
| CritPt | 20.9% | 0% |
| Chess Puzzles | 21% | 12% |
| LMArena Hard Prompts | 1480 | 1433 |
| Mystery Game Puzzles | 19% | 9% |
| DTBench | 93.6% | 80.3% |
| LMCA | 45.8% | 29.3% |
| Epoch Capabilities Index | 151.78 | 143.85 |
| NYT Connections (extended) | 74.3% | — |
| EBR-Bench | 9.5% | — |
| Surface Evolver Bench | 55.6% | — |
| ForecastBench | — | 59.7 |

## Math

- GLM-5.2: 55.7 (#43)
- Qwen3 235B-A22B: 50.4 (#57)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 86.4% | 86.7% |
| LMArena Math | 1482 | 1432 |
| FrontierMath (Tiers 1-3) | 59.2% | — |
| FrontierMath Tier 4 | 29.3% | — |
| MathArena Final-Answer Competitions | 67.6% | — |
| ProofBench | 35% | — |
| Omni-MATH | — | 71.8% |
| MATH Level 5 | — | 68.9% |
| FrontierMath (Feb 2025 set) | — | 8.5% |
| FrontierMath Tier 4 (v1) | — | 0% |

## Knowledge

- GLM-5.2: 57.1 (#40)
- Qwen3 235B-A22B: 49.6 (#73)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| GPQA Diamond | 91.9% | 80.1% |
| SimpleQA Verified | 34.2% | 40.4% |
| LMArena Expert | 1486 | 1463 |
| MMLU-Pro | — | 84.4% |
| Confabulations | — | 15.6% |
| Vectara Hallucination Rate | — | 9.3% |
| GPQA (HELM) | — | 72.7% |

## Multilingual

- GLM-5.2: 55.8 (#26)
- Qwen3 235B-A22B: 52.3 (#89)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Non-English | 1459 | 1409 |
| LMArena Chinese | 1519 | 1481 |
| LMArena French | 1479 | 1445 |
| LMArena German | 1468 | 1433 |
| LMArena Japanese | 1451 | 1399 |
| LMArena Korean | 1445 | 1391 |
| LMArena Russian | 1466 | 1411 |
| LMArena Spanish | 1477 | 1430 |

## Instruction Following

- GLM-5.2: 76.9 (#34)
- Qwen3 235B-A22B: 72.6 (#136)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Instruction Following | 1465 | 1408 |
| IFEval | — | 83.5% |

## Long Context

- GLM-5.2: 45.3 (#43)
- Qwen3 235B-A22B: 46.1 (#26)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Longer Query | 1479 | 1426 |
| Fiction.LiveBench | — | 75% |

## Writing & Preference

- GLM-5.2: 70.4 (#21)
- Qwen3 235B-A22B: 59.6 (#108)

| Benchmark | GLM-5.2 | Qwen3 235B-A22B |
|---|---|---|
| LMArena Text | 1470 | 1419 |
| LMArena Creative Writing | 1462 | 1384 |
| EQ-Bench Creative Writing | 1757 | 1366 |
| LMArena Multi-Turn | 1469 | 1432 |
| Short-Story Creative Writing | — | 83% |
| WildBench | — | 86.6% |
| EQ-Bench 4 | 1222 | — |

## FAQ

### Is GLM-5.2 better than Qwen3 235B-A22B?

GLM-5.2 is the stronger model overall, scoring 51.1 to 43.5 on the Noometry Index. Qwen3 235B-A22B costs 1.8× less per token, which makes it the better buy when GLM-5.2's lead doesn't matter for your workload.

### Which is cheaper, GLM-5.2 or Qwen3 235B-A22B?

Qwen3 235B-A22B is cheaper. It lists at $0.70 per million input tokens and $2.80 per million output tokens; GLM-5.2 lists at $1.40 and $4.40.

### Is GLM-5.2 or Qwen3 235B-A22B better for coding?

GLM-5.2 scores higher on coding benchmarks: 51.3 versus 44.3 in the Noometry coding category.

### Which has the bigger context window?

GLM-5.2 does, with 1M tokens against 131K.

### How many benchmarks do GLM-5.2 and Qwen3 235B-A22B share?

34 benchmarks have published results for both models. GLM-5.2 has 51 scored results on Noometry and Qwen3 235B-A22B has 49.