Model comparison
Gemma 3 27B vs GPT-4o mini
Gemma 3 27B is the stronger model overall, scoring 30.8 to 25.5 on the Noometry Index.
Last verified . 38 shared benchmarks.
GoogleSummary
- They share 38 benchmarks with published results for both. Gemma 3 27B scores higher in 8 categories and GPT-4o mini in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Gemma 3 27B leads 25.9 to 10.4.
- The biggest single-benchmark swing is MATH Level 5: 74% for Gemma 3 27B and 52.6% for GPT-4o mini.
- Gemma 3 27B is cheaper at $0.08 / $0.16 per million input/output tokens, against $0.15 / $0.60 for GPT-4o mini.
- Gemma 3 27B accepts more context: 131K tokens versus 128K.
- Gemma 3 27B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 27B | GPT-4o mini | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 30.8 | 25.5 |
| Released | 2025-03-11 | 2024-07-18 |
| Weights | Open | Proprietary |
| Context window | 131K | 128K |
| Max output | 8K | 16K |
| Input $ / M tokens | $0.08 | $0.15 |
| Output $ / M tokens | $0.16 | $0.60 |
| Results tracked | 43 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 3 27B: 22.5 (#334), GPT-4o mini: 22.0 (#335)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| Aider Polyglot | 4.9% | 3.6% |
| LiveBench Coding | 39.9% | 43.1% |
| LMArena Coding | 1322 | 1290 |
| SciCode | 21.2% | — |
| WeirdML | — | 11.8% |
| BigCodeBench Instruct | — | 46.1% |
| BigCodeBench Complete | — | 57.4% |
| HumanEval+ | — | 83.5% |
| MBPP+ | — | 72.2% |
Agentic & Tool Use GPT-4o mini leads
Gemma 3 27B: 25.1 (#110), GPT-4o mini: 27.5 (#101)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | 29.5% | — |
| BALROG | — | 17.4% |
Reasoning Gemma 3 27B leads
Gemma 3 27B: 16.7 (#301), GPT-4o mini: 8.7 (#347)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| Kagi LLM Benchmark | 40.4% | 28.8% |
| Chess Puzzles | 0% | 0% |
| LiveBench Reasoning | 43.8% | 32.8% |
| LMArena Hard Prompts | 1340 | 1267 |
| DTBench | 52.5% | 54.4% |
| LiveBench Data Analysis | 51.5% | 50% |
| LMCA | 12.3% | 10.4% |
| Epoch Capabilities Index | 130.04 | 126.56 |
| LiveBench | 50% | 41.3% |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 10.7% |
| CritPt | 0% | — |
| Mystery Game Puzzles | — | 12% |
| PIQA | — | 88.7% |
Math Gemma 3 27B leads
Gemma 3 27B: 25.9 (#265), GPT-4o mini: 10.4 (#314)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 6.9% |
| LiveBench Math | 55.4% | 36.3% |
| LMArena Math | 1312 | 1267 |
| MATH Level 5 | 74% | 52.6% |
| FrontierMath (Tiers 1-3) | — | 0.7% |
| Omni-MATH | — | 28% |
| GSM8K | — | 91.3% |
Knowledge Gemma 3 27B leads
Gemma 3 27B: 25.5 (#261), GPT-4o mini: 17.7 (#284)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| GPQA Diamond | 47.7% | 37.7% |
| Confabulations | 40.3% | 37.2% |
| LMArena Expert | 1304 | 1235 |
| SimpleQA Verified | — | 8.3% |
| MMLU-Pro | — | 60.3% |
| Vectara Hallucination Rate | 7.4% | — |
| GPQA (HELM) | — | 36.8% |
| BoolQ | — | 88.7% |
| MMLU | — | 81.8% |
Multimodal Gemma 3 27B leads
Gemma 3 27B: 32.6 (#100), GPT-4o mini: 25.9 (#122)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| LMArena Vision | 1164 | 1066 |
| GeoBench | 52% | 64% |
| Video-MME | — | 64.8% |
| VPCT | — | 34% |
Multilingual Gemma 3 27B leads
Gemma 3 27B: 46.9 (#155), GPT-4o mini: 42.0 (#199)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| LMArena Non-English | 1334 | 1266 |
| LMArena Chinese | 1346 | 1265 |
| LMArena French | 1368 | 1297 |
| LMArena German | 1362 | 1272 |
| LMArena Japanese | 1287 | 1216 |
| LMArena Korean | 1308 | 1195 |
| LMArena Russian | 1349 | 1275 |
| LMArena Spanish | 1349 | 1276 |
Instruction Following Gemma 3 27B leads
Gemma 3 27B: 70.6 (#160), GPT-4o mini: 61.9 (#239)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| LiveBench Instruction Following | 74.9% | 56.8% |
| LMArena Instruction Following | 1321 | 1258 |
| IFEval | — | 78.2% |
Long Context GPT-4o mini leads
Gemma 3 27B: 27.6 (#293), GPT-4o mini: 39.1 (#186)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| LMArena Longer Query | 1333 | 1289 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Gemma 3 27B leads
Gemma 3 27B: 52.5 (#168), GPT-4o mini: 39.5 (#248)
| Benchmark | Gemma 3 27B | GPT-4o mini |
|---|---|---|
| LMArena Text | 1358 | 1286 |
| LMArena Creative Writing | 1346 | 1268 |
| Short-Story Creative Writing | 79.9% | 67.2% |
| EQ-Bench Creative Writing | 1266 | 873 |
| LMArena Multi-Turn | 1345 | 1285 |
| LiveBench Language | 34.6% | 28.6% |
| WildBench | — | 79.1% |
Frequently asked questions
Is Gemma 3 27B better than GPT-4o mini?
Gemma 3 27B is the stronger model overall, scoring 30.8 to 25.5 on the Noometry Index.
Which is cheaper, Gemma 3 27B or GPT-4o mini?
Gemma 3 27B is cheaper. It lists at $0.08 per million input tokens and $0.16 per million output tokens; GPT-4o mini lists at $0.15 and $0.60.
Is Gemma 3 27B or GPT-4o mini better for coding?
They score almost the same on coding (22.5 vs 22.0); test both on your own repository before choosing.
Which has the bigger context window?
Gemma 3 27B does, with 131K tokens against 128K.
How many benchmarks do Gemma 3 27B and GPT-4o mini share?
38 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and GPT-4o mini has 60.