Model comparison
Gemma 2 27B vs Grok 3
Grok 3 is the stronger model overall, scoring 39.6 to 29.2 on the Noometry Index.
Last verified . 22 shared benchmarks.
GoogleSummary
- They share 22 benchmarks with published results for both. Gemma 2 27B scores higher in 1 category and Grok 3 in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok 3 leads 46.0 to 18.9.
- The biggest single-benchmark swing is MATH Level 5: 27.9% for Gemma 2 27B and 88.7% for Grok 3.
- Gemma 2 27B has downloadable open weights; the other is API-only.
Side by side
| Gemma 2 27B | Grok 3 | |
|---|---|---|
| Provider | xAI | |
| Noometry Index | 29.2 | 39.6 |
| Released | 2024-06-24 | 2025-04-09 |
| Weights | Open | Proprietary |
| Context window | 8K | — |
| Max output | 2K | — |
| Input $ / M tokens | $0.65 | — |
| Output $ / M tokens | $0.65 | — |
| Results tracked | 34 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 3 leads
Gemma 2 27B: 33.5 (#243), Grok 3: 41.5 (#113)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| LMArena Coding | 1211 | 1432 |
| Aider Polyglot | — | 53.3% |
| WeirdML | — | 37.2% |
| BigCodeBench Instruct | 42.8% | — |
| LiveBench Coding | 36% | — |
| BigCodeBench Complete | 52.5% | — |
Agentic & Tool Use Not comparable
Gemma 2 27B: —, Grok 3: 30.2 (#69)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| BALROG | — | 29.5% |
Reasoning Too close to call
Gemma 2 27B: 14.8 (#313), Grok 3: 13.8 (#322)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| LMArena Hard Prompts | 1199 | 1433 |
| Epoch Capabilities Index | 122.06 | 138.34 |
| ARC-AGI-2 | — | 0% |
| SimpleBench | — | 36.1% |
| Kagi LLM Benchmark | — | 61.3% |
| ARC-AGI-1 | — | 5.5% |
| LiveBench Reasoning | 28.1% | — |
| DTBench | 48% | — |
| LiveBench Data Analysis | 47.9% | — |
| LMCA | 7.1% | — |
| LiveBench | 38.2% | — |
Math Grok 3 leads
Gemma 2 27B: 10.0 (#306), Grok 3: 36.2 (#122)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.4% | 55.6% |
| LMArena Math | 1212 | 1390 |
| MATH Level 5 | 27.9% | 88.7% |
| Omni-MATH | — | 46.4% |
| LiveBench Math | 26.5% | — |
| FrontierMath (Feb 2025 set) | — | 3.8% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Grok 3 leads
Gemma 2 27B: 18.9 (#275), Grok 3: 46.0 (#76)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| GPQA Diamond | 36.5% | 75.8% |
| Confabulations | 27.1% | 14.2% |
| LMArena Expert | 1172 | 1420 |
| MMLU-Pro | — | 78.8% |
| Vectara Hallucination Rate | — | 5.8% |
| GPQA (HELM) | — | 65% |
| MMLU | 75.7% | — |
Multilingual Grok 3 leads
Gemma 2 27B: 38.8 (#221), Grok 3: 52.5 (#81)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| LMArena Non-English | 1217 | 1411 |
| LMArena Chinese | 1221 | 1448 |
| LMArena French | 1246 | 1460 |
| LMArena German | 1209 | 1431 |
| LMArena Japanese | 1175 | 1389 |
| LMArena Korean | 1174 | 1374 |
| LMArena Russian | 1234 | 1416 |
| LMArena Spanish | 1228 | 1417 |
Instruction Following Grok 3 leads
Gemma 2 27B: 60.7 (#244), Grok 3: 75.2 (#69)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| LMArena Instruction Following | 1206 | 1409 |
| LiveBench Instruction Following | 58.1% | — |
| IFEval | — | 88.4% |
Long Context Grok 3 leads
Gemma 2 27B: 37.3 (#213), Grok 3: 38.7 (#187)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| LMArena Longer Query | 1231 | 1439 |
| Fiction.LiveBench | — | 58.3% |
Writing & Preference Grok 3 leads
Gemma 2 27B: 44.4 (#220), Grok 3: 56.0 (#136)
| Benchmark | Gemma 2 27B | Grok 3 |
|---|---|---|
| LMArena Text | 1231 | 1426 |
| LMArena Creative Writing | 1241 | 1414 |
| LMArena Multi-Turn | 1224 | 1424 |
| Short-Story Creative Writing | — | 76.4% |
| EQ-Bench Creative Writing | — | 1186 |
| WildBench | — | 84.9% |
| LiveBench Language | 32.6% | — |
Frequently asked questions
Is Gemma 2 27B better than Grok 3?
Grok 3 is the stronger model overall, scoring 39.6 to 29.2 on the Noometry Index.
Is Gemma 2 27B or Grok 3 better for coding?
Grok 3 scores higher on coding benchmarks: 41.5 versus 33.5 in the Noometry coding category.
How many benchmarks do Gemma 2 27B and Grok 3 share?
22 benchmarks have published results for both models. Gemma 2 27B has 34 scored results on Noometry and Grok 3 has 40.