Model comparison
Gemma 2 27B vs GPT-4o mini
Gemma 2 27B is the stronger model overall, scoring 29.2 to 25.4 on the Noometry Index. GPT-4o mini costs 2.5× less per token, which makes it the better buy when Gemma 2 27B's lead doesn't matter for your workload.
Last verified . 34 shared benchmarks.
GoogleSummary
- They share 34 benchmarks with published results for both. Gemma 2 27B scores higher in 5 categories and GPT-4o mini in 3 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in coding, where Gemma 2 27B leads 33.5 to 21.6.
- The biggest single-benchmark swing is MATH Level 5: 27.9% for Gemma 2 27B and 52.6% for GPT-4o mini.
- GPT-4o mini is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.65 / $0.65 for Gemma 2 27B.
- GPT-4o mini accepts more context: 128K tokens versus 8K.
- Gemma 2 27B has downloadable open weights; the other is API-only.
Side by side
| Gemma 2 27B | GPT-4o mini | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 29.2 | 25.4 |
| Released | 2024-06-24 | 2024-07-18 |
| Weights | Open | Proprietary |
| Context window | 8K | 128K |
| Max output | 2K | 16K |
| Input $ / M tokens | $0.65 | $0.15 |
| Output $ / M tokens | $0.65 | $0.60 |
| Results tracked | 34 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 2 27B leads
Gemma 2 27B: 33.5 (#243), GPT-4o mini: 21.6 (#330)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| BigCodeBench Instruct | 42.8% | 46.1% |
| LiveBench Coding | 36% | 43.1% |
| LMArena Coding | 1211 | 1290 |
| BigCodeBench Complete | 52.5% | 57.4% |
| Aider Polyglot | — | 3.6% |
| WeirdML | — | 11.8% |
| HumanEval+ | — | 83.5% |
| MBPP+ | — | 72.2% |
Agentic & Tool Use Not comparable
Gemma 2 27B: —, GPT-4o mini: 27.2 (#95)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| BALROG | — | 17.4% |
Reasoning Gemma 2 27B leads
Gemma 2 27B: 14.8 (#313), GPT-4o mini: 8.7 (#343)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| LiveBench Reasoning | 28.1% | 32.8% |
| LMArena Hard Prompts | 1199 | 1267 |
| DTBench | 48% | 54.4% |
| LiveBench Data Analysis | 47.9% | 50% |
| LMCA | 7.1% | 10.4% |
| Epoch Capabilities Index | 122.06 | 126.56 |
| LiveBench | 38.2% | 41.3% |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 10.7% |
| Kagi LLM Benchmark | — | 28.8% |
| Chess Puzzles | — | 0% |
| Mystery Game Puzzles | — | 12% |
| PIQA | — | 88.7% |
Math Too close to call
Gemma 2 27B: 10.0 (#306), GPT-4o mini: 9.9 (#307)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.4% | 6.9% |
| LiveBench Math | 26.5% | 36.3% |
| LMArena Math | 1212 | 1267 |
| MATH Level 5 | 27.9% | 52.6% |
| FrontierMath (Tiers 1-3) | — | 0.7% |
| Omni-MATH | — | 28% |
| GSM8K | — | 91.3% |
Knowledge Gemma 2 27B leads
Gemma 2 27B: 18.9 (#275), GPT-4o mini: 17.7 (#279)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| GPQA Diamond | 36.5% | 37.7% |
| Confabulations | 27.1% | 37.2% |
| LMArena Expert | 1172 | 1235 |
| MMLU | 75.7% | 81.8% |
| SimpleQA Verified | — | 8.3% |
| MMLU-Pro | — | 60.3% |
| GPQA (HELM) | — | 36.8% |
| BoolQ | — | 88.7% |
Multimodal Not comparable
Gemma 2 27B: —, GPT-4o mini: 25.3 (#117)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| LMArena Vision | — | 1066 |
| Video-MME | — | 64.8% |
| GeoBench | — | 64% |
| VPCT | — | 34% |
Multilingual GPT-4o mini leads
Gemma 2 27B: 38.8 (#221), GPT-4o mini: 42.2 (#194)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| LMArena Non-English | 1217 | 1266 |
| LMArena Chinese | 1221 | 1264 |
| LMArena French | 1246 | 1296 |
| LMArena German | 1209 | 1272 |
| LMArena Japanese | 1175 | 1216 |
| LMArena Korean | 1174 | 1195 |
| LMArena Russian | 1234 | 1275 |
| LMArena Spanish | 1228 | 1276 |
Instruction Following GPT-4o mini leads
Gemma 2 27B: 60.7 (#244), GPT-4o mini: 62.1 (#234)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| LiveBench Instruction Following | 58.1% | 56.8% |
| LMArena Instruction Following | 1206 | 1259 |
| IFEval | — | 78.2% |
Long Context GPT-4o mini leads
Gemma 2 27B: 37.3 (#213), GPT-4o mini: 39.1 (#181)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| LMArena Longer Query | 1231 | 1289 |
Writing & Preference Gemma 2 27B leads
Gemma 2 27B: 44.4 (#220), GPT-4o mini: 39.7 (#243)
| Benchmark | Gemma 2 27B | GPT-4o mini |
|---|---|---|
| LMArena Text | 1231 | 1286 |
| LMArena Creative Writing | 1241 | 1268 |
| LMArena Multi-Turn | 1224 | 1285 |
| LiveBench Language | 32.6% | 28.6% |
| Short-Story Creative Writing | — | 67.2% |
| EQ-Bench Creative Writing | — | 873 |
| WildBench | — | 79.1% |
Frequently asked questions
Is Gemma 2 27B better than GPT-4o mini?
Gemma 2 27B is the stronger model overall, scoring 29.2 to 25.4 on the Noometry Index. GPT-4o mini costs 2.5× less per token, which makes it the better buy when Gemma 2 27B's lead doesn't matter for your workload.
Which is cheaper, Gemma 2 27B or GPT-4o mini?
GPT-4o mini is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Gemma 2 27B lists at $0.65 and $0.65.
Is Gemma 2 27B or GPT-4o mini better for coding?
Gemma 2 27B scores higher on coding benchmarks: 33.5 versus 21.6 in the Noometry coding category.
Which has the bigger context window?
GPT-4o mini does, with 128K tokens against 8K.
How many benchmarks do Gemma 2 27B and GPT-4o mini share?
34 benchmarks have published results for both models. Gemma 2 27B has 34 scored results on Noometry and GPT-4o mini has 60.