Model comparison
Gemma 3 12B vs Llama 3-8B
Gemma 3 12B is the stronger model overall, scoring 32.1 to 25.5 on the Noometry Index.
Last verified . 17 shared benchmarks.
Google
MetaSummary
- They share 17 benchmarks with published results for both. Gemma 3 12B scores higher in 8 categories and Llama 3-8B in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Gemma 3 12B leads 26.5 to 7.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 16.7% for Gemma 3 12B and 1.9% for Llama 3-8B.
Side by side
| Gemma 3 12B | Llama 3-8B | |
|---|---|---|
| Provider | Meta | |
| Noometry Index | 32.1 | 25.5 |
| Released | 2025-03-12 | 2024-04-18 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.05 | — |
| Output $ / M tokens | $0.15 | — |
| Results tracked | 24 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Gemma 3 12B: 31.7 (#280), Llama 3-8B: 31.0 (#289)
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| LMArena Coding | 1281 | 1152 |
| SciCode | 17.4% | — |
| BigCodeBench Instruct | — | 31.9% |
| BigCodeBench Complete | — | 36.9% |
| HumanEval+ | — | 56.7% |
| MBPP+ | — | 54.8% |
Agentic & Tool Use Not comparable
Gemma 3 12B: 25.5 (#108), Llama 3-8B: —
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 30.4% | — |
Reasoning Gemma 3 12B leads
Gemma 3 12B: 15.7 (#313), Llama 3-8B: 14.3 (#326)
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| Chess Puzzles | 0% | 0% |
| LMArena Hard Prompts | 1309 | 1133 |
| DTBench | 48.8% | 43.9% |
| Epoch Capabilities Index | 123.5 | 116.45 |
| CritPt | 0% | — |
| LMCA | 4.5% | — |
| Adversarial NLI | — | 57.3% |
| ForecastBench | — | 58.6 |
| WinoGrande | — | 75.7% |
Math Gemma 3 12B leads
Gemma 3 12B: 22.3 (#279), Llama 3-8B: 8.8 (#323)
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 16.7% | 1.9% |
| LMArena Math | 1307 | 1151 |
| MATH Level 5 | — | 6.1% |
Knowledge Gemma 3 12B leads
Gemma 3 12B: 26.5 (#257), Llama 3-8B: 7.8 (#308)
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| GPQA Diamond | 39.5% | 26.1% |
| LMArena Expert | 1248 | 1113 |
| Vectara Hallucination Rate | 4.4% | — |
| ARC (AI2) Challenge | — | 82.8% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 82.6% |
| TriviaQA | — | 67.7% |
Multimodal Not comparable
Gemma 3 12B: —, Llama 3-8B: —
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| MindCube | 46.7% | — |
Multilingual Gemma 3 12B leads
Gemma 3 12B: 45.7 (#165), Llama 3-8B: 30.8 (#261)
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| LMArena Non-English | 1318 | 1098 |
| LMArena German | 1370 | 1104 |
| LMArena Russian | 1335 | 1109 |
| LMArena Chinese | — | 1076 |
| LMArena French | — | 1159 |
| LMArena Japanese | — | 967 |
| LMArena Korean | — | 1004 |
| LMArena Spanish | — | 1173 |
Instruction Following Gemma 3 12B leads
Gemma 3 12B: 68.6 (#186), Llama 3-8B: 58.4 (#260)
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| LMArena Instruction Following | 1299 | 1127 |
Long Context Gemma 3 12B leads
Gemma 3 12B: 40.0 (#162), Llama 3-8B: 34.2 (#251)
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| LMArena Longer Query | 1317 | 1128 |
Writing & Preference Gemma 3 12B leads
Gemma 3 12B: 47.5 (#209), Llama 3-8B: 37.5 (#256)
| Benchmark | Gemma 3 12B | Llama 3-8B |
|---|---|---|
| LMArena Text | 1334 | 1166 |
| LMArena Creative Writing | 1331 | 1150 |
| LMArena Multi-Turn | 1334 | 1152 |
| EQ-Bench Creative Writing | 1126 | — |
Frequently asked questions
Is Gemma 3 12B better than Llama 3-8B?
Gemma 3 12B is the stronger model overall, scoring 32.1 to 25.5 on the Noometry Index.
Is Gemma 3 12B or Llama 3-8B better for coding?
They score almost the same on coding (31.7 vs 31.0); test both on your own repository before choosing.
How many benchmarks do Gemma 3 12B and Llama 3-8B share?
17 benchmarks have published results for both models. Gemma 3 12B has 24 scored results on Noometry and Llama 3-8B has 34.