Model comparison
Gemma 3 27B vs Gemma 3 4B
Gemma 3 27B is the stronger model overall, scoring 30.8 to 28.1 on the Noometry Index. Gemma 3 4B costs 2.0× less per token, which makes it the better buy when Gemma 3 27B's lead doesn't matter for your workload.
Last verified . 22 shared benchmarks.
GoogleSummary
- They share 22 benchmarks with published results for both. Gemma 3 27B scores higher in 7 categories and Gemma 3 4B in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Gemma 3 27B leads 25.5 to 11.8.
- The biggest single-benchmark swing is GPQA Diamond: 47.7% for Gemma 3 27B and 23.2% for Gemma 3 4B.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.08 / $0.16 for Gemma 3 27B.
Side by side
| Gemma 3 27B | Gemma 3 4B | |
|---|---|---|
| Provider | ||
| Noometry Index | 30.8 | 28.1 |
| Released | 2025-03-11 | 2025-03-12 |
| Weights | Open | Open |
| Context window | 131K | 131K |
| Max output | 8K | 4K |
| Input $ / M tokens | $0.08 | $0.04 |
| Output $ / M tokens | $0.16 | $0.08 |
| Results tracked | 43 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 3 4B leads
Gemma 3 27B: 22.5 (#334), Gemma 3 4B: 35.9 (#215)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| LMArena Coding | 1322 | 1230 |
| Aider Polyglot | 4.9% | — |
| SciCode | 21.2% | — |
| LiveBench Coding | 39.9% | — |
Agentic & Tool Use Gemma 3 27B leads
Gemma 3 27B: 25.1 (#110), Gemma 3 4B: 20.9 (#142)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 29.5% | 19.6% |
Reasoning Gemma 3 27B leads
Gemma 3 27B: 16.7 (#301), Gemma 3 4B: 13.2 (#335)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| Kagi LLM Benchmark | 40.4% | 25.2% |
| Chess Puzzles | 0% | 0% |
| LMArena Hard Prompts | 1340 | 1253 |
| DTBench | 52.5% | 50.9% |
| LMCA | 12.3% | 2.8% |
| Epoch Capabilities Index | 130.04 | 116.02 |
| CritPt | 0% | — |
| LiveBench Reasoning | 43.8% | — |
| LiveBench Data Analysis | 51.5% | — |
| LiveBench | 50% | — |
Math Gemma 3 27B leads
Gemma 3 27B: 25.9 (#265), Gemma 3 4B: 16.8 (#292)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 7.5% |
| LMArena Math | 1312 | 1239 |
| LiveBench Math | 55.4% | — |
| MATH Level 5 | 74% | — |
Knowledge Gemma 3 27B leads
Gemma 3 27B: 25.5 (#261), Gemma 3 4B: 11.8 (#299)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| GPQA Diamond | 47.7% | 23.2% |
| Vectara Hallucination Rate | 7.4% | 6.4% |
| LMArena Expert | 1304 | 1223 |
| Confabulations | 40.3% | — |
Multimodal Not comparable
Gemma 3 27B: 32.6 (#100), Gemma 3 4B: —
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| LMArena Vision | 1164 | — |
| GeoBench | 52% | — |
Multilingual Gemma 3 27B leads
Gemma 3 27B: 46.9 (#155), Gemma 3 4B: 42.5 (#194)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| LMArena Non-English | 1334 | 1273 |
| LMArena German | 1362 | 1281 |
| LMArena Russian | 1349 | 1294 |
| LMArena Chinese | 1346 | — |
| LMArena French | 1368 | — |
| LMArena Japanese | 1287 | — |
| LMArena Korean | 1308 | — |
| LMArena Spanish | 1349 | — |
Instruction Following Gemma 3 27B leads
Gemma 3 27B: 70.6 (#160), Gemma 3 4B: 65.2 (#225)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| LMArena Instruction Following | 1321 | 1239 |
| LiveBench Instruction Following | 74.9% | — |
Long Context Gemma 3 4B leads
Gemma 3 27B: 27.6 (#293), Gemma 3 4B: 38.7 (#194)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| LMArena Longer Query | 1333 | 1273 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Gemma 3 27B leads
Gemma 3 27B: 52.5 (#168), Gemma 3 4B: 42.0 (#239)
| Benchmark | Gemma 3 27B | Gemma 3 4B |
|---|---|---|
| LMArena Text | 1358 | 1291 |
| LMArena Creative Writing | 1346 | 1271 |
| EQ-Bench Creative Writing | 1266 | 1068 |
| LMArena Multi-Turn | 1345 | 1255 |
| Short-Story Creative Writing | 79.9% | — |
| LiveBench Language | 34.6% | — |
Frequently asked questions
Is Gemma 3 27B better than Gemma 3 4B?
Gemma 3 27B is the stronger model overall, scoring 30.8 to 28.1 on the Noometry Index. Gemma 3 4B costs 2.0× less per token, which makes it the better buy when Gemma 3 27B's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 27B or Gemma 3 4B?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; Gemma 3 27B lists at $0.08 and $0.16.
Is Gemma 3 27B or Gemma 3 4B better for coding?
Gemma 3 4B scores higher on coding benchmarks: 35.9 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 Gemma 3 4B share?
22 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and Gemma 3 4B has 22.