Model comparison
Gemma 3 27B vs Qwen2.5 72B Instruct
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.9 to 30.8 on the Noometry Index. Gemma 3 27B costs 24× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Last verified . 24 shared benchmarks.
Google
Alibaba (Qwen)Summary
- They share 24 benchmarks with published results for both. Gemma 3 27B scores higher in 5 categories and Qwen2.5 72B Instruct in 4 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where Qwen2.5 72B Instruct leads 38.9 to 27.6.
- The biggest single-benchmark swing is Confabulations: 40.3% for Gemma 3 27B and 19.1% for Qwen2.5 72B Instruct.
- Gemma 3 27B is cheaper at $0.08 / $0.16 per million input/output tokens, against $1.40 / $5.60 for Qwen2.5 72B Instruct.
Side by side
| Gemma 3 27B | Qwen2.5 72B Instruct | |
|---|---|---|
| Provider | Alibaba (Qwen) | |
| Noometry Index | 30.8 | 31.9 |
| Released | 2025-03-11 | 2024-09 |
| Weights | Open | Open |
| Context window | 131K | 131K |
| Max output | 8K | 8K |
| Input $ / M tokens | $0.08 | $1.40 |
| Output $ / M tokens | $0.16 | $5.60 |
| Results tracked | 43 | 43 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen2.5 72B Instruct leads
Gemma 3 27B: 22.5 (#334), Qwen2.5 72B Instruct: 33.2 (#260)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Coding | 1322 | 1292 |
| Aider Polyglot | 4.9% | — |
| SciCode | 21.2% | — |
| WeirdML | — | 16% |
| BigCodeBench Instruct | — | 45.8% |
| LiveBench Coding | 39.9% | — |
| BigCodeBench Complete | — | 55.9% |
Agentic & Tool Use Gemma 3 27B leads
Gemma 3 27B: 25.1 (#110), Qwen2.5 72B Instruct: 22.1 (#133)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| Berkeley Function Calling Leaderboard | 29.5% | — |
| TheAgentCompany | — | 5.7% |
| BALROG | — | 16.2% |
| METR Time Horizons | — | 35.8% |
Reasoning Qwen2.5 72B Instruct leads
Gemma 3 27B: 16.7 (#301), Qwen2.5 72B Instruct: 22.3 (#199)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Hard Prompts | 1340 | 1271 |
| DTBench | 52.5% | 62.9% |
| LMCA | 12.3% | 13.4% |
| Epoch Capabilities Index | 130.04 | 129 |
| Kagi LLM Benchmark | 40.4% | — |
| CritPt | 0% | — |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | 43.8% | — |
| LiveBench Data Analysis | 51.5% | — |
| BIG-Bench Hard | — | 79.8% |
| ForecastBench | — | 57.5 |
| HellaSwag | — | 84.8% |
| LiveBench | 50% | — |
| PIQA | — | 82.6% |
| WinoGrande | — | 82.3% |
Math Gemma 3 27B leads
Gemma 3 27B: 25.9 (#265), Qwen2.5 72B Instruct: 19.3 (#287)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 8.1% |
| LMArena Math | 1312 | 1283 |
| MATH Level 5 | 74% | 63.2% |
| Omni-MATH | — | 33% |
| LiveBench Math | 55.4% | — |
Knowledge Qwen2.5 72B Instruct leads
Gemma 3 27B: 25.5 (#261), Qwen2.5 72B Instruct: 27.0 (#253)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| GPQA Diamond | 47.7% | 49.1% |
| Confabulations | 40.3% | 19.1% |
| LMArena Expert | 1304 | 1245 |
| MMLU-Pro | — | 63.1% |
| Vectara Hallucination Rate | 7.4% | — |
| GPQA (HELM) | — | 42.6% |
| ARC (AI2) Challenge | — | 94.5% |
| MMLU | — | 85.3% |
| TriviaQA | — | 71.9% |
Multimodal Not comparable
Gemma 3 27B: 32.6 (#100), Qwen2.5 72B Instruct: —
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Vision | 1164 | — |
| GeoBench | 52% | — |
Multilingual Gemma 3 27B leads
Gemma 3 27B: 46.9 (#155), Qwen2.5 72B Instruct: 41.0 (#213)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Non-English | 1334 | 1252 |
| LMArena Chinese | 1346 | 1272 |
| LMArena French | 1368 | 1280 |
| LMArena German | 1362 | 1234 |
| LMArena Japanese | 1287 | 1180 |
| LMArena Korean | 1308 | 1188 |
| LMArena Russian | 1349 | 1264 |
| LMArena Spanish | 1349 | 1256 |
Instruction Following Gemma 3 27B leads
Gemma 3 27B: 70.6 (#160), Qwen2.5 72B Instruct: 65.5 (#221)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Instruction Following | 1321 | 1254 |
| LiveBench Instruction Following | 74.9% | — |
| IFEval | — | 80.6% |
Long Context Qwen2.5 72B Instruct leads
Gemma 3 27B: 27.6 (#293), Qwen2.5 72B Instruct: 38.9 (#188)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Longer Query | 1333 | 1282 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Gemma 3 27B leads
Gemma 3 27B: 52.5 (#168), Qwen2.5 72B Instruct: 46.7 (#215)
| Benchmark | Gemma 3 27B | Qwen2.5 72B Instruct |
|---|---|---|
| LMArena Text | 1358 | 1269 |
| LMArena Creative Writing | 1346 | 1221 |
| LMArena Multi-Turn | 1345 | 1272 |
| Short-Story Creative Writing | 79.9% | — |
| EQ-Bench Creative Writing | 1266 | — |
| WildBench | — | 80.2% |
| LiveBench Language | 34.6% | — |
Frequently asked questions
Is Gemma 3 27B better than Qwen2.5 72B Instruct?
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.9 to 30.8 on the Noometry Index. Gemma 3 27B costs 24× less per token, which makes it the better buy when Qwen2.5 72B Instruct's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 27B or Qwen2.5 72B Instruct?
Gemma 3 27B is cheaper. It lists at $0.08 per million input tokens and $0.16 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Is Gemma 3 27B or Qwen2.5 72B Instruct better for coding?
Qwen2.5 72B Instruct scores higher on coding benchmarks: 33.2 versus 22.5 in the Noometry coding category.
Which has the bigger context window?
Both accept 131K tokens.
How many benchmarks do Gemma 3 27B and Qwen2.5 72B Instruct share?
24 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and Qwen2.5 72B Instruct has 43.