# Gemini 3 Deep Think vs GLM-5.2

> GLM-5.2 has enough public results to be ranked (#38); Gemini 3 Deep Think does not yet, so treat this comparison as directional.

- Canonical page: https://noometry.com/compare/gemini-3-deep-think-vs-glm-5-2
- Last updated: 2026-10-10
- Shared benchmarks: 3

## Summary

- They share 3 benchmarks with published results for both. Gemini 3 Deep Think scores higher in 1 category and GLM-5.2 in 0 categories; one gap is clear of the uncertainty.
- The widest gap is in reasoning, where Gemini 3 Deep Think leads 70.8 to 42.1.
- The biggest single-benchmark swing is ARC-AGI-2: 84.6% for Gemini 3 Deep Think and 22.8% for GLM-5.2.
- GLM-5.2 has downloadable open weights; the other is API-only.

## Snapshot

| | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| Provider | Google | Z.ai (Zhipu) |
| Noometry Index | 58.5 | 50.9 |
| Rank | — | 38 |
| Context | — | 1M |
| Input $/M | — | $1.40 |
| Output $/M | — | $4.40 |
| Weights | Proprietary | Open |

## Coding

- Gemini 3 Deep Think: —
- GLM-5.2: 50.9 (#37)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| SWE-bench Verified | — | 78.7% |
| DeepSWE | — | 43.8% |
| FrontierCode | — | 24.5% |
| LMArena WebDev | — | 1604 |
| SciCode | — | 50.5% |
| WeirdML | — | 70.1% |
| LMArena Coding | — | 1484 |
| ALE-Bench | — | 1,047 |

## Agentic & Tool Use

- Gemini 3 Deep Think: —
- GLM-5.2: 31.6 (#63)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 31.7% |
| GBAEval | — | 0% |
| Vending-Bench 2 | — | 8,314 |

## Reasoning

- Gemini 3 Deep Think: 70.8 (#11)
- GLM-5.2: 42.1 (#46)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| ARC-AGI-2 | 84.6% | 22.8% |
| ARC-AGI-1 | 96% | 77% |
| CritPt | 25.7% | 20.9% |
| SimpleBench | — | 58.8% |
| Kagi LLM Benchmark | — | 62.6% |
| NYT Connections (extended) | — | 74.3% |
| Chess Puzzles | — | 21% |
| EBR-Bench | — | 9.5% |
| LMArena Hard Prompts | — | 1480 |
| Mystery Game Puzzles | — | 19% |
| DTBench | — | 93.6% |
| LMCA | — | 45.8% |
| Surface Evolver Bench | — | 55.6% |
| Epoch Capabilities Index | — | 151.76 |

## Math

- Gemini 3 Deep Think: —
- GLM-5.2: 55.4 (#35)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 59.2% |
| FrontierMath Tier 4 | — | 29.3% |
| MathArena Final-Answer Competitions | — | 67.6% |
| OTIS Mock AIME 2024-2025 | — | 86.4% |
| ProofBench | — | 35% |
| LMArena Math | — | 1482 |

## Knowledge

- Gemini 3 Deep Think: —
- GLM-5.2: 57.1 (#36)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| GPQA Diamond | — | 91.9% |
| SimpleQA Verified | — | 34.2% |
| LMArena Expert | — | 1484 |

## Multilingual

- Gemini 3 Deep Think: —
- GLM-5.2: 56.0 (#25)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| LMArena Non-English | — | 1459 |
| LMArena Chinese | — | 1518 |
| LMArena French | — | 1476 |
| LMArena German | — | 1470 |
| LMArena Japanese | — | 1451 |
| LMArena Korean | — | 1445 |
| LMArena Russian | — | 1467 |
| LMArena Spanish | — | 1477 |

## Instruction Following

- Gemini 3 Deep Think: —
- GLM-5.2: 77.1 (#32)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| LMArena Instruction Following | — | 1465 |

## Long Context

- Gemini 3 Deep Think: —
- GLM-5.2: 45.3 (#43)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| LMArena Longer Query | — | 1479 |

## Writing & Preference

- Gemini 3 Deep Think: —
- GLM-5.2: 70.7 (#20)

| Benchmark | Gemini 3 Deep Think | GLM-5.2 |
|---|---|---|
| LMArena Text | — | 1470 |
| LMArena Creative Writing | — | 1465 |
| EQ-Bench Creative Writing | — | 1757 |
| EQ-Bench 4 | — | 1222 |
| LMArena Multi-Turn | — | 1468 |

## FAQ

### Is Gemini 3 Deep Think better than GLM-5.2?

GLM-5.2 has enough public results to be ranked (#38); Gemini 3 Deep Think does not yet, so treat this comparison as directional.

### How many benchmarks do Gemini 3 Deep Think and GLM-5.2 share?

3 benchmarks have published results for both models. Gemini 3 Deep Think has 3 scored results on Noometry and GLM-5.2 has 50.