Model comparison
Gemma 2 27B vs GLM-5.3
GLM-5.3 is the stronger model overall, scoring 54.8 to 29.4 on the Noometry Index. Gemma 2 27B costs 3.3× less per token, which makes it the better buy when GLM-5.3's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
GoogleSummary
- They share 22 benchmarks with published results for both. Gemma 2 27B scores higher in 0 categories and GLM-5.3 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GLM-5.3 leads 62.3 to 10.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.4% for Gemma 2 27B and 91.1% for GLM-5.3.
- Gemma 2 27B is cheaper at $0.65 / $0.65 per million input/output tokens, against $1.40 / $4.40 for GLM-5.3.
- GLM-5.3 accepts more context: 1M tokens versus 8K.
Side by side
| Gemma 2 27B | GLM-5.3 | |
|---|---|---|
| Provider | Z.ai (Zhipu) | |
| Noometry Index | 29.4 | 54.8 |
| Released | 2024-06-24 | 2026-08-14 |
| Weights | Open | Open |
| Context window | 8K | 1M |
| Max output | 2K | 131K |
| Input $ / M tokens | $0.65 | $1.40 |
| Output $ / M tokens | $0.65 | $4.40 |
| Results tracked | 34 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GLM-5.3 leads
Gemma 2 27B: 34.1 (#246), GLM-5.3: 59.5 (#14)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| LMArena Coding | 1211 | 1496 |
| DeepSWE | — | 69% |
| FrontierCode | — | 40.1% |
| CursorBench | — | 42.6% |
| LMArena WebDev | — | 1622 |
| FrontierSWE | — | 30.2% |
| SciCode | — | 59% |
| WeirdML | — | 75.4% |
| BigCodeBench Instruct | 42.8% | — |
| LiveBench Coding | 36% | — |
| BigCodeBench Complete | 52.5% | — |
| ALE-Bench | — | 1,317 |
Agentic & Tool Use Not comparable
Gemma 2 27B: —, GLM-5.3: 36.4 (#38)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| APEX-Agents | — | 56.6% |
| Vending-Bench 2 | — | 8,164 |
Reasoning GLM-5.3 leads
Gemma 2 27B: 15.3 (#315), GLM-5.3: 46.1 (#46)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| LMArena Hard Prompts | 1198 | 1489 |
| DTBench | 48% | 87.7% |
| LMCA | 7.1% | 55.5% |
| Epoch Capabilities Index | 122.08 | 155.61 |
| NYT Connections (extended) | — | 74.2% |
| CritPt | — | 19.1% |
| Chess Puzzles | — | 21% |
| LiveBench Reasoning | 28.1% | — |
| Mystery Game Puzzles | — | 33% |
| LiveBench Data Analysis | 47.9% | — |
| Bench to the Future 3 | — | 0.15 |
| LiveBench | 38.2% | — |
Math GLM-5.3 leads
Gemma 2 27B: 10.7 (#311), GLM-5.3: 62.3 (#33)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.4% | 91.1% |
| LMArena Math | 1212 | 1489 |
| FrontierMath (Tiers 1-3) | — | 68.8% |
| FrontierMath Tier 4 | — | 29.3% |
| ProofBench | — | 49% |
| LiveBench Math | 26.5% | — |
| MATH Level 5 | 27.9% | — |
Knowledge GLM-5.3 leads
Gemma 2 27B: 19.0 (#280), GLM-5.3: 58.3 (#37)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| GPQA Diamond | 36.5% | 90.9% |
| LMArena Expert | 1172 | 1516 |
| SimpleQA Verified | — | 41% |
| Confabulations | 27.1% | — |
| MMLU | 75.7% | — |
Multilingual GLM-5.3 leads
Gemma 2 27B: 38.6 (#226), GLM-5.3: 55.7 (#28)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| LMArena Non-English | 1217 | 1457 |
| LMArena Chinese | 1221 | 1528 |
| LMArena French | 1247 | 1499 |
| LMArena German | 1209 | 1499 |
| LMArena Japanese | 1175 | 1453 |
| LMArena Korean | 1174 | 1472 |
| LMArena Russian | 1234 | 1463 |
| LMArena Spanish | 1228 | 1460 |
Instruction Following GLM-5.3 leads
Gemma 2 27B: 60.5 (#249), GLM-5.3: 77.5 (#23)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| LMArena Instruction Following | 1206 | 1477 |
| LiveBench Instruction Following | 58.1% | — |
Long Context GLM-5.3 leads
Gemma 2 27B: 37.3 (#218), GLM-5.3: 45.4 (#41)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| LMArena Longer Query | 1231 | 1482 |
Writing & Preference GLM-5.3 leads
Gemma 2 27B: 44.2 (#225), GLM-5.3: 75.7 (#6)
| Benchmark | Gemma 2 27B | GLM-5.3 |
|---|---|---|
| LMArena Text | 1231 | 1471 |
| LMArena Creative Writing | 1241 | 1457 |
| LMArena Multi-Turn | 1224 | 1472 |
| EQ-Bench Creative Writing | — | 2075 |
| LiveBench Language | 32.6% | — |
Frequently asked questions
Is Gemma 2 27B better than GLM-5.3?
GLM-5.3 is the stronger model overall, scoring 54.8 to 29.4 on the Noometry Index. Gemma 2 27B costs 3.3× less per token, which makes it the better buy when GLM-5.3's lead doesn't matter for your workload.
Which is cheaper, Gemma 2 27B or GLM-5.3?
Gemma 2 27B is cheaper. It lists at $0.65 per million input tokens and $0.65 per million output tokens; GLM-5.3 lists at $1.40 and $4.40.
Is Gemma 2 27B or GLM-5.3 better for coding?
GLM-5.3 scores higher on coding benchmarks: 59.5 versus 34.1 in the Noometry coding category.
Which has the bigger context window?
GLM-5.3 does, with 1M tokens against 8K.
How many benchmarks do Gemma 2 27B and GLM-5.3 share?
22 benchmarks have published results for both models. Gemma 2 27B has 34 scored results on Noometry and GLM-5.3 has 42.