Model comparison
Gemma 3 27B vs Llama 3-70B
Gemma 3 27B is the stronger model overall, scoring 30.8 to 28.8 on the Noometry Index.
Last verified . 23 shared benchmarks.
Google
MetaSummary
- They share 23 benchmarks with published results for both. Gemma 3 27B scores higher in 6 categories and Llama 3-70B in 3 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in coding, where Llama 3-70B leads 35.8 to 22.5.
- The biggest single-benchmark swing is MATH Level 5: 74% for Gemma 3 27B and 22.6% for Llama 3-70B.
Side by side
| Gemma 3 27B | Llama 3-70B | |
|---|---|---|
| Provider | Meta | |
| Noometry Index | 30.8 | 28.8 |
| Released | 2025-03-11 | 2024-04-18 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.08 | — |
| Output $ / M tokens | $0.16 | — |
| Results tracked | 43 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 3-70B leads
Gemma 3 27B: 22.5 (#334), Llama 3-70B: 35.8 (#218)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| LMArena Coding | 1322 | 1206 |
| Aider Polyglot | 4.9% | — |
| SciCode | 21.2% | — |
| BigCodeBench Instruct | — | 43.6% |
| LiveBench Coding | 39.9% | — |
| BigCodeBench Complete | — | 54.5% |
| HumanEval+ | — | 72% |
| MBPP+ | — | 69% |
Agentic & Tool Use Gemma 3 27B leads
Gemma 3 27B: 25.1 (#110), Llama 3-70B: 21.1 (#139)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 29.5% | — |
| Cybench | — | 5% |
Reasoning Llama 3-70B leads
Gemma 3 27B: 16.7 (#301), Llama 3-70B: 18.0 (#288)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| Kagi LLM Benchmark | 40.4% | 35.1% |
| LMArena Hard Prompts | 1340 | 1195 |
| DTBench | 52.5% | 54.2% |
| Epoch Capabilities Index | 130.04 | 122.93 |
| CritPt | 0% | — |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | 43.8% | — |
| LiveBench Data Analysis | 51.5% | — |
| LMCA | 12.3% | — |
| ForecastBench | — | 57.1 |
| LiveBench | 50% | — |
| WinoGrande | — | 83.5% |
Math Gemma 3 27B leads
Gemma 3 27B: 25.9 (#265), Llama 3-70B: 12.8 (#305)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 4.3% |
| LMArena Math | 1312 | 1218 |
| MATH Level 5 | 74% | 22.6% |
| LiveBench Math | 55.4% | — |
Knowledge Gemma 3 27B leads
Gemma 3 27B: 25.5 (#261), Llama 3-70B: 20.8 (#277)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| GPQA Diamond | 47.7% | 40.6% |
| LMArena Expert | 1304 | 1149 |
| Confabulations | 40.3% | — |
| Vectara Hallucination Rate | 7.4% | — |
| MMLU | — | 79.3% |
Multimodal Not comparable
Gemma 3 27B: 32.6 (#100), Llama 3-70B: —
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| LMArena Vision | 1164 | — |
| GeoBench | 52% | — |
Multilingual Gemma 3 27B leads
Gemma 3 27B: 46.9 (#155), Llama 3-70B: 33.6 (#251)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| LMArena Non-English | 1334 | 1142 |
| LMArena Chinese | 1346 | 1114 |
| LMArena French | 1368 | 1232 |
| LMArena German | 1362 | 1169 |
| LMArena Japanese | 1287 | 1017 |
| LMArena Korean | 1308 | 1017 |
| LMArena Russian | 1349 | 1159 |
| LMArena Spanish | 1349 | 1241 |
Instruction Following Gemma 3 27B leads
Gemma 3 27B: 70.6 (#160), Llama 3-70B: 62.5 (#238)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| LMArena Instruction Following | 1321 | 1194 |
| LiveBench Instruction Following | 74.9% | — |
Long Context Llama 3-70B leads
Gemma 3 27B: 27.6 (#293), Llama 3-70B: 35.6 (#240)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| LMArena Longer Query | 1333 | 1174 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Gemma 3 27B leads
Gemma 3 27B: 52.5 (#168), Llama 3-70B: 42.8 (#231)
| Benchmark | Gemma 3 27B | Llama 3-70B |
|---|---|---|
| LMArena Text | 1358 | 1221 |
| LMArena Creative Writing | 1346 | 1210 |
| LMArena Multi-Turn | 1345 | 1223 |
| Short-Story Creative Writing | 79.9% | — |
| EQ-Bench Creative Writing | 1266 | — |
| LiveBench Language | 34.6% | — |
Frequently asked questions
Is Gemma 3 27B better than Llama 3-70B?
Gemma 3 27B is the stronger model overall, scoring 30.8 to 28.8 on the Noometry Index.
Is Gemma 3 27B or Llama 3-70B better for coding?
Llama 3-70B scores higher on coding benchmarks: 35.8 versus 22.5 in the Noometry coding category.
How many benchmarks do Gemma 3 27B and Llama 3-70B share?
23 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and Llama 3-70B has 31.