Model comparison
Gemma 3 27B vs Grok-2 (Dec 2024)
Grok-2 (Dec 2024) is the stronger model overall, scoring 33.7 to 30.8 on the Noometry Index.
Last verified . 31 shared benchmarks.
GoogleSummary
- They share 31 benchmarks with published results for both. Gemma 3 27B scores higher in 4 categories and Grok-2 (Dec 2024) in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in long context, where Grok-2 (Dec 2024) leads 38.8 to 27.6.
- The biggest single-benchmark swing is Confabulations: 40.3% for Gemma 3 27B and 20.1% for Grok-2 (Dec 2024).
- Gemma 3 27B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 27B | Grok-2 (Dec 2024) | |
|---|---|---|
| Provider | xAI | |
| Noometry Index | 30.8 | 33.7 |
| Released | 2025-03-11 | 2024-08-13 |
| Weights | Open | Proprietary |
| Context window | 131K | — |
| Max output | 8K | — |
| Input $ / M tokens | $0.08 | — |
| Output $ / M tokens | $0.16 | — |
| Results tracked | 43 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok-2 (Dec 2024) leads
Gemma 3 27B: 22.5 (#334), Grok-2 (Dec 2024): 33.3 (#258)
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| LiveBench Coding | 39.9% | 46.4% |
| LMArena Coding | 1322 | 1287 |
| Aider Polyglot | 4.9% | — |
| SciCode | 21.2% | — |
| WeirdML | — | 22.2% |
Agentic & Tool Use Not comparable
Gemma 3 27B: 25.1 (#110), Grok-2 (Dec 2024): —
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| Berkeley Function Calling Leaderboard | 29.5% | — |
Reasoning Too close to call
Gemma 3 27B: 16.7 (#301), Grok-2 (Dec 2024): 16.9 (#299)
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| LiveBench Reasoning | 43.8% | 54.8% |
| LMArena Hard Prompts | 1340 | 1272 |
| DTBench | 52.5% | 65.2% |
| LiveBench Data Analysis | 51.5% | 54.5% |
| Epoch Capabilities Index | 130.04 | 130.48 |
| LiveBench | 50% | 54.3% |
| SimpleBench | — | 22.7% |
| Kagi LLM Benchmark | 40.4% | — |
| CritPt | 0% | — |
| Chess Puzzles | 0% | — |
| LMCA | 12.3% | — |
Math Gemma 3 27B leads
Gemma 3 27B: 25.9 (#265), Grok-2 (Dec 2024): 20.8 (#284)
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 11.5% |
| LiveBench Math | 55.4% | 54.9% |
| LMArena Math | 1312 | 1283 |
| MATH Level 5 | 74% | 63.5% |
| FrontierMath (Feb 2025 set) | — | 0.7% |
Knowledge Grok-2 (Dec 2024) leads
Gemma 3 27B: 25.5 (#261), Grok-2 (Dec 2024): 29.8 (#233)
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| GPQA Diamond | 47.7% | 53.8% |
| Confabulations | 40.3% | 20.1% |
| LMArena Expert | 1304 | 1254 |
| Vectara Hallucination Rate | 7.4% | — |
Multimodal Not comparable
Gemma 3 27B: 32.6 (#100), Grok-2 (Dec 2024): —
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| LMArena Vision | 1164 | — |
| GeoBench | 52% | — |
Multilingual Gemma 3 27B leads
Gemma 3 27B: 46.9 (#155), Grok-2 (Dec 2024): 43.1 (#188)
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| LMArena Non-English | 1334 | 1282 |
| LMArena Chinese | 1346 | 1289 |
| LMArena French | 1368 | 1318 |
| LMArena German | 1362 | 1287 |
| LMArena Japanese | 1287 | 1244 |
| LMArena Korean | 1308 | 1237 |
| LMArena Russian | 1349 | 1286 |
| LMArena Spanish | 1349 | 1281 |
Instruction Following Gemma 3 27B leads
Gemma 3 27B: 70.6 (#160), Grok-2 (Dec 2024): 66.9 (#202)
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| LiveBench Instruction Following | 74.9% | 69.6% |
| LMArena Instruction Following | 1321 | 1270 |
Long Context Grok-2 (Dec 2024) leads
Gemma 3 27B: 27.6 (#293), Grok-2 (Dec 2024): 38.8 (#190)
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| LMArena Longer Query | 1333 | 1276 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Gemma 3 27B leads
Gemma 3 27B: 52.5 (#168), Grok-2 (Dec 2024): 48.6 (#198)
| Benchmark | Gemma 3 27B | Grok-2 (Dec 2024) |
|---|---|---|
| LMArena Text | 1358 | 1305 |
| LMArena Creative Writing | 1346 | 1284 |
| Short-Story Creative Writing | 79.9% | 63.6% |
| LMArena Multi-Turn | 1345 | 1290 |
| LiveBench Language | 34.6% | 45.6% |
| EQ-Bench Creative Writing | 1266 | — |
Frequently asked questions
Is Gemma 3 27B better than Grok-2 (Dec 2024)?
Grok-2 (Dec 2024) is the stronger model overall, scoring 33.7 to 30.8 on the Noometry Index.
Is Gemma 3 27B or Grok-2 (Dec 2024) better for coding?
Grok-2 (Dec 2024) scores higher on coding benchmarks: 33.3 versus 22.5 in the Noometry coding category.
How many benchmarks do Gemma 3 27B and Grok-2 (Dec 2024) share?
31 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and Grok-2 (Dec 2024) has 34.