Model comparison
Gemma 2 27B vs GPT-4.1
GPT-4.1 is the stronger model overall, scoring 35.9 to 29.4 on the Noometry Index. Gemma 2 27B costs 5.4× less per token, which makes it the better buy when GPT-4.1's lead doesn't matter for your workload.
Last verified . 23 shared benchmarks.
GoogleSummary
- They share 23 benchmarks with published results for both. Gemma 2 27B scores higher in 1 category and GPT-4.1 in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-4.1 leads 37.1 to 19.0.
- The biggest single-benchmark swing is MATH Level 5: 27.9% for Gemma 2 27B and 83% for GPT-4.1.
- Gemma 2 27B is cheaper at $0.65 / $0.65 per million input/output tokens, against $2 / $8 for GPT-4.1.
- GPT-4.1 accepts more context: 1.05M tokens versus 8K.
- Gemma 2 27B has downloadable open weights; the other is API-only.
Side by side
| Gemma 2 27B | GPT-4.1 | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 29.4 | 35.9 |
| Released | 2024-06-24 | 2025-04-14 |
| Weights | Open | Proprietary |
| Context window | 8K | 1.05M |
| Max output | 2K | 33K |
| Input $ / M tokens | $0.65 | $2 |
| Output $ / M tokens | $0.65 | $8 |
| Results tracked | 34 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 2 27B: 34.1 (#246), GPT-4.1: 34.4 (#238)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| LMArena Coding | 1211 | 1391 |
| SWE-bench Verified | — | 48.5% |
| SWE-bench Verified (bash only) | — | 39.6% |
| Aider Polyglot | — | 52.4% |
| WeirdML | — | 39% |
| BigCodeBench Instruct | 42.8% | — |
| LiveBench Coding | 36% | — |
| BigCodeBench Complete | 52.5% | — |
| CadEval | — | 42% |
| ALE-Bench | — | 558.1 |
Agentic & Tool Use Not comparable
Gemma 2 27B: —, GPT-4.1: 34.7 (#43)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 54% |
Reasoning Gemma 2 27B leads
Gemma 2 27B: 15.3 (#315), GPT-4.1: 11.7 (#339)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| LMArena Hard Prompts | 1198 | 1384 |
| DTBench | 48% | 68.3% |
| LMCA | 7.1% | 25.6% |
| Epoch Capabilities Index | 122.08 | 136.78 |
| ARC-AGI-2 | — | 0.4% |
| SimpleBench | — | 27% |
| Kagi LLM Benchmark | — | 52.3% |
| ARC-AGI-1 | — | 5.5% |
| Chess Puzzles | — | 6% |
| EnigmaEval | — | 2.2% |
| LiveBench Reasoning | 28.1% | — |
| LiveBench Data Analysis | 47.9% | — |
| ForecastBench | — | 61.5 |
| LiveBench | 38.2% | — |
Math GPT-4.1 leads
Gemma 2 27B: 10.7 (#311), GPT-4.1: 22.3 (#280)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.4% | 38.3% |
| LMArena Math | 1212 | 1370 |
| MATH Level 5 | 27.9% | 83% |
| FrontierMath (Tiers 1-3) | — | 6% |
| Omni-MATH | — | 47.1% |
| LiveBench Math | 26.5% | — |
| FrontierMath (Feb 2025 set) | — | 5.5% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge GPT-4.1 leads
Gemma 2 27B: 19.0 (#280), GPT-4.1: 37.1 (#160)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| GPQA Diamond | 36.5% | 66.9% |
| LMArena Expert | 1172 | 1364 |
| Humanity's Last Exam | — | 5.4% |
| SimpleQA Verified | — | 31.1% |
| MMLU-Pro | — | 81.1% |
| Confabulations | 27.1% | — |
| Vectara Hallucination Rate | — | 5.6% |
| GPQA (HELM) | — | 65.9% |
| MMLU | 75.7% | — |
Multimodal Not comparable
Gemma 2 27B: —, GPT-4.1: 38.2 (#67)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| LMArena Vision | — | 1211 |
| GeoBench | — | 72% |
Multilingual GPT-4.1 leads
Gemma 2 27B: 38.6 (#226), GPT-4.1: 49.4 (#133)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| LMArena Non-English | 1217 | 1370 |
| LMArena Chinese | 1221 | 1382 |
| LMArena French | 1247 | 1382 |
| LMArena German | 1209 | 1381 |
| LMArena Japanese | 1175 | 1319 |
| LMArena Korean | 1174 | 1339 |
| LMArena Russian | 1234 | 1377 |
| LMArena Spanish | 1228 | 1376 |
Instruction Following GPT-4.1 leads
Gemma 2 27B: 60.5 (#249), GPT-4.1: 71.3 (#153)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| LMArena Instruction Following | 1206 | 1367 |
| LiveBench Instruction Following | 58.1% | — |
| IFEval | — | 83.8% |
Long Context GPT-4.1 leads
Gemma 2 27B: 37.3 (#218), GPT-4.1: 40.0 (#163)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| LMArena Longer Query | 1231 | 1385 |
| Fiction.LiveBench | — | 63.9% |
Writing & Preference GPT-4.1 leads
Gemma 2 27B: 44.2 (#225), GPT-4.1: 57.6 (#125)
| Benchmark | Gemma 2 27B | GPT-4.1 |
|---|---|---|
| LMArena Text | 1231 | 1383 |
| LMArena Creative Writing | 1241 | 1363 |
| LMArena Multi-Turn | 1224 | 1398 |
| EQ-Bench Creative Writing | — | 1420 |
| WildBench | — | 85.4% |
| LiveBench Language | 32.6% | — |
Frequently asked questions
Is Gemma 2 27B better than GPT-4.1?
GPT-4.1 is the stronger model overall, scoring 35.9 to 29.4 on the Noometry Index. Gemma 2 27B costs 5.4× less per token, which makes it the better buy when GPT-4.1's lead doesn't matter for your workload.
Which is cheaper, Gemma 2 27B or GPT-4.1?
Gemma 2 27B is cheaper. It lists at $0.65 per million input tokens and $0.65 per million output tokens; GPT-4.1 lists at $2 and $8.
Is Gemma 2 27B or GPT-4.1 better for coding?
They score almost the same on coding (34.1 vs 34.4); test both on your own repository before choosing.
Which has the bigger context window?
GPT-4.1 does, with 1.05M tokens against 8K.
How many benchmarks do Gemma 2 27B and GPT-4.1 share?
23 benchmarks have published results for both models. Gemma 2 27B has 34 scored results on Noometry and GPT-4.1 has 52.