Model comparison
Gemma 3 27B vs GPT-5.4 nano
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 30.5 on the Noometry Index. Gemma 3 27B costs 4.6× less per token, which makes it the better buy when GPT-5.4 nano's lead doesn't matter for your workload.
Last verified . 28 shared benchmarks.
GoogleSummary
- They share 28 benchmarks with published results for both. Gemma 3 27B scores higher in 0 categories and GPT-5.4 nano in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in coding, where GPT-5.4 nano leads 43.1 to 22.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 22.5% for Gemma 3 27B and 87.8% for GPT-5.4 nano.
- Gemma 3 27B is cheaper at $0.08 / $0.16 per million input/output tokens, against $0.20 / $1.25 for GPT-5.4 nano.
- GPT-5.4 nano accepts more context: 400K tokens versus 131K.
- Gemma 3 27B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 27B | GPT-5.4 nano | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 30.5 | 41.9 |
| Released | 2025-03-11 | 2026-03-17 |
| Weights | Open | Proprietary |
| Context window | 131K | 400K |
| Max output | 8K | 128K |
| Input $ / M tokens | $0.08 | $0.20 |
| Output $ / M tokens | $0.16 | $1.25 |
| Results tracked | 43 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 nano leads
Gemma 3 27B: 22.1 (#329), GPT-5.4 nano: 43.1 (#82)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| SciCode | 21.2% | 46.9% |
| LMArena Coding | 1322 | 1404 |
| Aider Polyglot | 4.9% | — |
| WeirdML | — | 49.2% |
| LiveBench Coding | 39.9% | — |
| ALE-Bench | — | 1,005 |
Agentic & Tool Use Not comparable
Gemma 3 27B: 24.9 (#105), GPT-5.4 nano: —
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| Berkeley Function Calling Leaderboard | 29.5% | — |
Reasoning GPT-5.4 nano leads
Gemma 3 27B: 16.3 (#297), GPT-5.4 nano: 23.8 (#159)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| Kagi LLM Benchmark | 40.4% | 39.7% |
| CritPt | 0% | 9.3% |
| Chess Puzzles | 0% | 30% |
| LMArena Hard Prompts | 1340 | 1380 |
| DTBench | 52.5% | 80.3% |
| LMCA | 12.3% | 36.9% |
| Epoch Capabilities Index | 130.04 | 145.78 |
| ARC-AGI-2 | — | 5.7% |
| ARC-AGI-1 | — | 51.5% |
| LiveBench Reasoning | 43.8% | — |
| Mystery Game Puzzles | — | 9% |
| LiveBench Data Analysis | 51.5% | — |
| ForecastBench | — | 57.3 |
| LiveBench | 50% | — |
Math GPT-5.4 nano leads
Gemma 3 27B: 24.4 (#261), GPT-5.4 nano: 40.7 (#76)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 87.8% |
| LMArena Math | 1312 | 1406 |
| FrontierMath (Tiers 1-3) | — | 44.9% |
| FrontierMath Tier 4 | — | 12.2% |
| ProofBench | — | 5% |
| LiveBench Math | 55.4% | — |
| MATH Level 5 | 74% | — |
| FrontierMath (Feb 2025 set) | — | 25.9% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge GPT-5.4 nano leads
Gemma 3 27B: 25.4 (#256), GPT-5.4 nano: 41.8 (#94)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| GPQA Diamond | 47.7% | 78.5% |
| Vectara Hallucination Rate | 7.4% | 3.1% |
| LMArena Expert | 1303 | 1397 |
| SimpleQA Verified | — | 11.7% |
| Confabulations | 40.3% | — |
Multimodal GPT-5.4 nano leads
Gemma 3 27B: 32.0 (#94), GPT-5.4 nano: 36.0 (#73)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| LMArena Vision | 1165 | 1196 |
| GeoBench | 52% | — |
Multilingual GPT-5.4 nano leads
Gemma 3 27B: 47.0 (#150), GPT-5.4 nano: 48.8 (#136)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| LMArena Non-English | 1334 | 1359 |
| LMArena Chinese | 1347 | 1391 |
| LMArena French | 1368 | 1396 |
| LMArena German | 1362 | 1368 |
| LMArena Japanese | 1287 | 1342 |
| LMArena Korean | 1308 | 1321 |
| LMArena Russian | 1349 | 1363 |
| LMArena Spanish | 1348 | 1368 |
Instruction Following GPT-5.4 nano leads
Gemma 3 27B: 70.8 (#155), GPT-5.4 nano: 72.1 (#139)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| LMArena Instruction Following | 1321 | 1362 |
| LiveBench Instruction Following | 74.9% | — |
Long Context GPT-5.4 nano leads
Gemma 3 27B: 27.6 (#288), GPT-5.4 nano: 41.6 (#131)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| LMArena Longer Query | 1333 | 1365 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference GPT-5.4 nano leads
Gemma 3 27B: 52.7 (#161), GPT-5.4 nano: 55.9 (#137)
| Benchmark | Gemma 3 27B | GPT-5.4 nano |
|---|---|---|
| LMArena Text | 1358 | 1372 |
| LMArena Creative Writing | 1346 | 1313 |
| LMArena Multi-Turn | 1344 | 1382 |
| Short-Story Creative Writing | 79.9% | — |
| EQ-Bench Creative Writing | 1266 | — |
| LiveBench Language | 34.6% | — |
Frequently asked questions
Is Gemma 3 27B better than GPT-5.4 nano?
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 30.5 on the Noometry Index. Gemma 3 27B costs 4.6× less per token, which makes it the better buy when GPT-5.4 nano's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 27B or GPT-5.4 nano?
Gemma 3 27B is cheaper. It lists at $0.08 per million input tokens and $0.16 per million output tokens; GPT-5.4 nano lists at $0.20 and $1.25.
Is Gemma 3 27B or GPT-5.4 nano better for coding?
GPT-5.4 nano scores higher on coding benchmarks: 43.1 versus 22.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 nano does, with 400K tokens against 131K.
How many benchmarks do Gemma 3 27B and GPT-5.4 nano share?
28 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and GPT-5.4 nano has 40.