Model comparison
Gemma 4 31B IT vs GLM-5.2
GLM-5.2 is the stronger model overall, scoring 51.1 to 43.5 on the Noometry Index. Gemma 4 31B IT costs 14× less per token, which makes it the better buy when GLM-5.2's lead doesn't matter for your workload.
Last verified . 31 shared benchmarks.
GoogleSummary
- They share 31 benchmarks with published results for both. Gemma 4 31B IT scores higher in 0 categories and GLM-5.2 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GLM-5.2 leads 57.1 to 37.9.
- The biggest single-benchmark swing is Surface Evolver Bench: 30.6% for Gemma 4 31B IT and 55.6% for GLM-5.2.
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $1.40 / $4.40 for GLM-5.2.
- GLM-5.2 accepts more context: 1M tokens versus 262K.
Side by side
| Gemma 4 31B IT | GLM-5.2 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | |
| Noometry Index | 43.5 | 51.1 |
| Released | 2026-04-02 | 2026-06-13 |
| Weights | Open | Open |
| Context window | 262K | 1M |
| Max output | 33K | 131K |
| Input $ / M tokens | $0.09 | $1.40 |
| Output $ / M tokens | $0.34 | $4.40 |
| Results tracked | 35 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.2 leads
Gemma 4 31B IT: 42.3 (#108), GLM-5.2: 51.3 (#41)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| LMArena WebDev | 1366 | 1603 |
| SciCode | 43.4% | 50.5% |
| WeirdML | 52.3% | 70.1% |
| LMArena Coding | 1459 | 1485 |
| ALE-Bench | 925.5 | 1,047 |
| SWE-bench Verified | — | 78.7% |
| DeepSWE | — | 43.8% |
| FrontierCode | — | 24.5% |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, GLM-5.2: 32.4 (#63)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| APEX-Agents | — | 45.2% |
| τ²-bench Banking | — | 37.1% |
| PostTrainBench | — | 31.7% |
| GBAEval | — | 0% |
| Vending-Bench 2 | — | 8,314 |
Reasoning GLM-5.2 leads
Gemma 4 31B IT: 27.2 (#122), GLM-5.2: 42.3 (#52)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| Kagi LLM Benchmark | 63.5% | 62.6% |
| NYT Connections (extended) | 70.6% | 74.3% |
| CritPt | 1.4% | 20.9% |
| Chess Puzzles | 5% | 21% |
| LMArena Hard Prompts | 1448 | 1480 |
| DTBench | 82.7% | 93.6% |
| LMCA | 39.3% | 45.8% |
| Surface Evolver Bench | 30.6% | 55.6% |
| Epoch Capabilities Index | 142.74 | 151.78 |
| ARC-AGI-2 | — | 22.8% |
| SimpleBench | — | 58.8% |
| ARC-AGI-1 | — | 77% |
| Thematic Generalization | 53% | — |
| EBR-Bench | — | 9.5% |
| Mystery Game Puzzles | — | 19% |
Math GLM-5.2 leads
Gemma 4 31B IT: 43.2 (#81), GLM-5.2: 55.7 (#43)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.3% | 86.4% |
| LMArena Math | 1465 | 1482 |
| FrontierMath (Tiers 1-3) | — | 59.2% |
| FrontierMath Tier 4 | — | 29.3% |
| MathArena Final-Answer Competitions | — | 67.6% |
| ProofBench | — | 35% |
Knowledge GLM-5.2 leads
Gemma 4 31B IT: 37.9 (#151), GLM-5.2: 57.1 (#40)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| GPQA Diamond | 75.8% | 91.9% |
| SimpleQA Verified | 10.4% | 34.2% |
| LMArena Expert | 1465 | 1486 |
| Vectara Hallucination Rate | 7.4% | — |
Multimodal Not comparable
Gemma 4 31B IT: 41.6 (#34), GLM-5.2: —
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| LMArena Vision | 1277 | — |
| LMArena Document | 1425 | — |
Multilingual GLM-5.2 leads
Gemma 4 31B IT: 53.8 (#57), GLM-5.2: 55.8 (#26)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| LMArena Non-English | 1431 | 1459 |
| LMArena Chinese | 1476 | 1519 |
| LMArena French | 1435 | 1479 |
| LMArena Russian | 1460 | 1466 |
| LMArena Spanish | 1444 | 1477 |
| LMArena German | — | 1468 |
| LMArena Japanese | — | 1451 |
| LMArena Korean | — | 1445 |
Instruction Following GLM-5.2 leads
Gemma 4 31B IT: 75.5 (#61), GLM-5.2: 76.9 (#34)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| LMArena Instruction Following | 1433 | 1465 |
Long Context GLM-5.2 leads
Gemma 4 31B IT: 44.2 (#71), GLM-5.2: 45.3 (#43)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| LMArena Longer Query | 1446 | 1479 |
Writing & Preference GLM-5.2 leads
Gemma 4 31B IT: 60.5 (#96), GLM-5.2: 70.4 (#21)
| Benchmark | Gemma 4 31B IT | GLM-5.2 |
|---|---|---|
| LMArena Text | 1443 | 1470 |
| LMArena Creative Writing | 1415 | 1462 |
| EQ-Bench Creative Writing | 1368 | 1757 |
| EQ-Bench 4 | 1120 | 1222 |
| LMArena Multi-Turn | 1452 | 1469 |
Frequently asked questions
Is Gemma 4 31B IT better than GLM-5.2?
GLM-5.2 is the stronger model overall, scoring 51.1 to 43.5 on the Noometry Index. Gemma 4 31B IT costs 14× less per token, which makes it the better buy when GLM-5.2's lead doesn't matter for your workload.
Which is cheaper, Gemma 4 31B IT or GLM-5.2?
Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; GLM-5.2 lists at $1.40 and $4.40.
Is Gemma 4 31B IT or GLM-5.2 better for coding?
GLM-5.2 scores higher on coding benchmarks: 51.3 versus 42.3 in the Noometry coding category.
Which has the bigger context window?
GLM-5.2 does, with 1M tokens against 262K.
How many benchmarks do Gemma 4 31B IT and GLM-5.2 share?
31 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and GLM-5.2 has 51.