Model comparison
Gemma 2 27B vs Qwen3.8 Max
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 29.4 on the Noometry Index. Gemma 2 27B costs 4.6× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
Google
Alibaba (Qwen)Summary
- They share 22 benchmarks with published results for both. Gemma 2 27B scores higher in 0 categories and Qwen3.8 Max in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 10.7.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.4% for Gemma 2 27B and 100% for Qwen3.8 Max.
- Gemma 2 27B is cheaper at $0.65 / $0.65 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
- Qwen3.8 Max accepts more context: 1M tokens versus 8K.
- Gemma 2 27B has downloadable open weights; the other is API-only.
Side by side
| Gemma 2 27B | Qwen3.8 Max | |
|---|---|---|
| Provider | Alibaba (Qwen) | |
| Noometry Index | 29.4 | 56.8 |
| Released | 2024-06-24 | 2026-08-02 |
| Weights | Open | Proprietary |
| Context window | 8K | 1M |
| Max output | 2K | 131K |
| Input $ / M tokens | $0.65 | $2 |
| Output $ / M tokens | $0.65 | $6 |
| Results tracked | 34 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 Max leads
Gemma 2 27B: 34.1 (#246), Qwen3.8 Max: 53.5 (#29)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| LMArena Coding | 1211 | 1502 |
| DeepSWE | — | 57.5% |
| LMArena WebDev | — | 1674 |
| FrontierSWE | — | 17.8% |
| SciCode | — | 53.2% |
| BigCodeBench Instruct | 42.8% | — |
| LiveBench Coding | 36% | — |
| BigCodeBench Complete | 52.5% | — |
Agentic & Tool Use Not comparable
Gemma 2 27B: —, Qwen3.8 Max: 45.4 (#14)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | — | 63.3% |
| τ²-bench Banking | — | 55.1% |
| GDP.pdf | — | 23.2% |
Reasoning Qwen3.8 Max leads
Gemma 2 27B: 15.3 (#315), Qwen3.8 Max: 54.4 (#26)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| LMArena Hard Prompts | 1198 | 1496 |
| DTBench | 48% | 92% |
| LMCA | 7.1% | 46.2% |
| Epoch Capabilities Index | 122.08 | 156.41 |
| NYT Connections (extended) | — | 88.3% |
| CritPt | — | 20% |
| Chess Puzzles | — | 40% |
| LiveBench Reasoning | 28.1% | — |
| Mystery Game Puzzles | — | 38% |
| LiveBench Data Analysis | 47.9% | — |
| LiveBench | 38.2% | — |
Math Qwen3.8 Max leads
Gemma 2 27B: 10.7 (#311), Qwen3.8 Max: 73.2 (#20)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.4% | 100% |
| LMArena Math | 1212 | 1499 |
| FrontierMath (Tiers 1-3) | — | 74.7% |
| FrontierMath Tier 4 | — | 46.3% |
| ProofBench | — | 58% |
| LiveBench Math | 26.5% | — |
| MATH Level 5 | 27.9% | — |
Knowledge Qwen3.8 Max leads
Gemma 2 27B: 19.0 (#280), Qwen3.8 Max: 61.7 (#27)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 36.5% | 92.7% |
| LMArena Expert | 1172 | 1507 |
| SimpleQA Verified | — | 47.3% |
| Confabulations | 27.1% | — |
| MMLU | 75.7% | — |
Multimodal Not comparable
Gemma 2 27B: —, Qwen3.8 Max: 37.2 (#75)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | — | 1314 |
| Furniture Assembly | — | 20% |
Multilingual Qwen3.8 Max leads
Gemma 2 27B: 38.6 (#226), Qwen3.8 Max: 56.7 (#18)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1217 | 1472 |
| LMArena Chinese | 1221 | 1538 |
| LMArena French | 1247 | 1503 |
| LMArena German | 1209 | 1483 |
| LMArena Japanese | 1175 | 1467 |
| LMArena Korean | 1174 | 1461 |
| LMArena Russian | 1234 | 1481 |
| LMArena Spanish | 1228 | 1492 |
Instruction Following Qwen3.8 Max leads
Gemma 2 27B: 60.5 (#249), Qwen3.8 Max: 77.6 (#17)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1206 | 1479 |
| LiveBench Instruction Following | 58.1% | — |
Long Context Qwen3.8 Max leads
Gemma 2 27B: 37.3 (#218), Qwen3.8 Max: 45.6 (#31)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1231 | 1489 |
Writing & Preference Qwen3.8 Max leads
Gemma 2 27B: 44.2 (#225), Qwen3.8 Max: 67.1 (#30)
| Benchmark | Gemma 2 27B | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1231 | 1483 |
| LMArena Creative Writing | 1241 | 1479 |
| LMArena Multi-Turn | 1224 | 1489 |
| LiveBench Language | 32.6% | — |
Frequently asked questions
Is Gemma 2 27B better than Qwen3.8 Max?
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 29.4 on the Noometry Index. Gemma 2 27B costs 4.6× less per token, which makes it the better buy when Qwen3.8 Max's lead doesn't matter for your workload.
Which is cheaper, Gemma 2 27B or Qwen3.8 Max?
Gemma 2 27B is cheaper. It lists at $0.65 per million input tokens and $0.65 per million output tokens; Qwen3.8 Max lists at $2 and $6.
Is Gemma 2 27B or Qwen3.8 Max better for coding?
Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 34.1 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 Max does, with 1M tokens against 8K.
How many benchmarks do Gemma 2 27B and Qwen3.8 Max share?
22 benchmarks have published results for both models. Gemma 2 27B has 34 scored results on Noometry and Qwen3.8 Max has 39.