Model comparison
Gemma 3 27B vs o1
o1 is the stronger model overall, scoring 40.9 to 30.8 on the Noometry Index. Gemma 3 27B costs 263× less per token, which makes it the better buy when o1's lead doesn't matter for your workload.
Last verified . 37 shared benchmarks.
GoogleSummary
- They share 37 benchmarks with published results for both. Gemma 3 27B scores higher in 1 category and o1 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in coding, where o1 leads 46.1 to 22.5.
- The biggest single-benchmark swing is Aider Polyglot: 4.9% for Gemma 3 27B and 61.7% for o1.
- Gemma 3 27B is cheaper at $0.08 / $0.16 per million input/output tokens, against $15 / $60 for o1.
- o1 accepts more context: 200K tokens versus 131K.
- Gemma 3 27B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 27B | o1 | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 30.8 | 40.9 |
| Released | 2025-03-11 | 2024-09-12 |
| Weights | Open | Proprietary |
| Context window | 131K | 200K |
| Max output | 8K | 100K |
| Input $ / M tokens | $0.08 | $15 |
| Output $ / M tokens | $0.16 | $60 |
| Results tracked | 43 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o1 leads
Gemma 3 27B: 22.5 (#334), o1: 46.1 (#70)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| Aider Polyglot | 4.9% | 61.7% |
| LiveBench Coding | 39.9% | 69.7% |
| LMArena Coding | 1322 | 1367 |
| SciCode | 21.2% | — |
| WeirdML | — | 47.6% |
| CadEval | — | 56% |
| HumanEval+ | — | 89% |
| MBPP+ | — | 80.2% |
Agentic & Tool Use Too close to call
Gemma 3 27B: 25.1 (#110), o1: 24.6 (#117)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 29.5% | — |
| Cybench | — | 10% |
| METR Time Horizons | — | 51.1% |
Reasoning o1 leads
Gemma 3 27B: 16.7 (#301), o1: 27.9 (#111)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| Chess Puzzles | 0% | 15% |
| LiveBench Reasoning | 43.8% | 91.6% |
| LMArena Hard Prompts | 1340 | 1371 |
| DTBench | 52.5% | 74.7% |
| LiveBench Data Analysis | 51.5% | 65.5% |
| LMCA | 12.3% | 22.3% |
| Epoch Capabilities Index | 130.04 | 141.91 |
| LiveBench | 50% | 75.7% |
| SimpleBench | — | 41.7% |
| Kagi LLM Benchmark | 40.4% | — |
| ARC-AGI-1 | — | 30.7% |
| CritPt | 0% | — |
| EnigmaEval | — | 5.7% |
Math o1 leads
Gemma 3 27B: 25.9 (#265), o1: 36.1 (#175)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 73.3% |
| LiveBench Math | 55.4% | 80.3% |
| LMArena Math | 1312 | 1388 |
| MATH Level 5 | 74% | 94.7% |
| FrontierMath (Tiers 1-3) | — | 14.7% |
| FrontierMath (Feb 2025 set) | — | 9.3% |
Knowledge o1 leads
Gemma 3 27B: 25.5 (#261), o1: 41.5 (#110)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| GPQA Diamond | 47.7% | 76.8% |
| Confabulations | 40.3% | 11.7% |
| LMArena Expert | 1304 | 1361 |
| Humanity's Last Exam | — | 8% |
| SimpleQA Verified | — | 41.1% |
| Vectara Hallucination Rate | 7.4% | — |
Multimodal o1 leads
Gemma 3 27B: 32.6 (#100), o1: 34.2 (#93)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| LMArena Vision | 1164 | 1168 |
| GeoBench | 52% | 80% |
| VPCT | — | 37% |
| SpatialViz-Bench | — | 41.4% |
Multilingual o1 leads
Gemma 3 27B: 46.9 (#155), o1: 48.6 (#142)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| LMArena Non-English | 1334 | 1358 |
| LMArena Chinese | 1346 | 1394 |
| LMArena French | 1368 | 1344 |
| LMArena German | 1362 | 1337 |
| LMArena Japanese | 1287 | 1346 |
| LMArena Korean | 1308 | 1396 |
| LMArena Russian | 1349 | 1356 |
| LMArena Spanish | 1349 | 1345 |
Instruction Following o1 leads
Gemma 3 27B: 70.6 (#160), o1: 74.8 (#86)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| LiveBench Instruction Following | 74.9% | 81.5% |
| LMArena Instruction Following | 1321 | 1367 |
Long Context o1 leads
Gemma 3 27B: 27.6 (#293), o1: 50.3 (#9)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| Fiction.LiveBench | 33.3% | 83.3% |
| LMArena Longer Query | 1333 | 1378 |
Writing & Preference o1 leads
Gemma 3 27B: 52.5 (#168), o1: 55.6 (#144)
| Benchmark | Gemma 3 27B | o1 |
|---|---|---|
| LMArena Text | 1358 | 1366 |
| LMArena Creative Writing | 1346 | 1348 |
| Short-Story Creative Writing | 79.9% | 70.2% |
| LMArena Multi-Turn | 1345 | 1369 |
| LiveBench Language | 34.6% | 65.4% |
| EQ-Bench Creative Writing | 1266 | — |
Frequently asked questions
Is Gemma 3 27B better than o1?
o1 is the stronger model overall, scoring 40.9 to 30.8 on the Noometry Index. Gemma 3 27B costs 263× less per token, which makes it the better buy when o1's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 27B or o1?
Gemma 3 27B is cheaper. It lists at $0.08 per million input tokens and $0.16 per million output tokens; o1 lists at $15 and $60.
Is Gemma 3 27B or o1 better for coding?
o1 scores higher on coding benchmarks: 46.1 versus 22.5 in the Noometry coding category.
Which has the bigger context window?
o1 does, with 200K tokens against 131K.
How many benchmarks do Gemma 3 27B and o1 share?
37 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and o1 has 52.