Model comparison
Gemma 3 27B vs gpt-oss-20b
gpt-oss-20b is the stronger model overall, scoring 32.5 to 30.8 on the Noometry Index.
Last verified . 26 shared benchmarks.
GoogleSummary
- They share 26 benchmarks with published results for both. Gemma 3 27B scores higher in 4 categories and gpt-oss-20b in 5 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Gemma 3 27B leads 52.5 to 35.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 22.5% for Gemma 3 27B and 65.3% for gpt-oss-20b.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.08 / $0.16 for Gemma 3 27B.
Side by side
| Gemma 3 27B | gpt-oss-20b | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 30.8 | 32.5 |
| Released | 2025-03-11 | 2025-08-05 |
| Weights | Open | Open |
| Context window | 131K | 131K |
| Max output | 8K | 16K |
| Input $ / M tokens | $0.08 | $0.018 |
| Output $ / M tokens | $0.16 | $0.09 |
| Results tracked | 43 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding gpt-oss-20b leads
Gemma 3 27B: 22.5 (#334), gpt-oss-20b: 37.6 (#192)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| SciCode | 21.2% | 34.4% |
| LMArena Coding | 1322 | 1306 |
| Aider Polyglot | 4.9% | — |
| WeirdML | — | 40.9% |
| LiveBench Coding | 39.9% | — |
| ALE-Bench | — | 566.05 |
Agentic & Tool Use Gemma 3 27B leads
Gemma 3 27B: 25.1 (#110), gpt-oss-20b: 9.3 (#154)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| Terminal-Bench | — | 3.4% |
| Berkeley Function Calling Leaderboard | 29.5% | — |
Reasoning gpt-oss-20b leads
Gemma 3 27B: 16.7 (#301), gpt-oss-20b: 19.3 (#261)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| Kagi LLM Benchmark | 40.4% | 53.2% |
| CritPt | 0% | 1.4% |
| Chess Puzzles | 0% | 4% |
| LMArena Hard Prompts | 1340 | 1274 |
| DTBench | 52.5% | 68% |
| LMCA | 12.3% | 14.5% |
| Epoch Capabilities Index | 130.04 | 137.82 |
| LiveBench Reasoning | 43.8% | — |
| LiveBench Data Analysis | 51.5% | — |
| LiveBench | 50% | — |
Math gpt-oss-20b leads
Gemma 3 27B: 25.9 (#265), gpt-oss-20b: 39.4 (#103)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 22.5% | 65.3% |
| LMArena Math | 1312 | 1317 |
| Omni-MATH | — | 56.5% |
| LiveBench Math | 55.4% | — |
| MATH Level 5 | 74% | — |
Knowledge gpt-oss-20b leads
Gemma 3 27B: 25.5 (#261), gpt-oss-20b: 34.6 (#195)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| GPQA Diamond | 47.7% | 60.8% |
| LMArena Expert | 1304 | 1258 |
| MMLU-Pro | — | 74% |
| Confabulations | 40.3% | — |
| Vectara Hallucination Rate | 7.4% | — |
| GPQA (HELM) | — | 59.4% |
Multimodal Not comparable
Gemma 3 27B: 32.6 (#100), gpt-oss-20b: —
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| LMArena Vision | 1164 | — |
| GeoBench | 52% | — |
Multilingual Gemma 3 27B leads
Gemma 3 27B: 46.9 (#155), gpt-oss-20b: 42.2 (#197)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| LMArena Non-English | 1334 | 1268 |
| LMArena Chinese | 1346 | 1314 |
| LMArena German | 1362 | 1255 |
| LMArena Japanese | 1287 | 1244 |
| LMArena Korean | 1308 | 1236 |
| LMArena Russian | 1349 | 1278 |
| LMArena Spanish | 1349 | 1267 |
| LMArena French | 1368 | — |
Instruction Following Gemma 3 27B leads
Gemma 3 27B: 70.6 (#160), gpt-oss-20b: 61.8 (#240)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| LMArena Instruction Following | 1321 | 1236 |
| LiveBench Instruction Following | 74.9% | — |
| IFEval | — | 73.2% |
Long Context gpt-oss-20b leads
Gemma 3 27B: 27.6 (#293), gpt-oss-20b: 37.9 (#209)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| LMArena Longer Query | 1333 | 1250 |
| Fiction.LiveBench | 33.3% | — |
Writing & Preference Gemma 3 27B leads
Gemma 3 27B: 52.5 (#168), gpt-oss-20b: 35.5 (#265)
| Benchmark | Gemma 3 27B | gpt-oss-20b |
|---|---|---|
| LMArena Text | 1358 | 1287 |
| LMArena Creative Writing | 1346 | 1201 |
| EQ-Bench Creative Writing | 1266 | 666 |
| LMArena Multi-Turn | 1345 | 1268 |
| Short-Story Creative Writing | 79.9% | — |
| WildBench | — | 73.7% |
| LiveBench Language | 34.6% | — |
Frequently asked questions
Is Gemma 3 27B better than gpt-oss-20b?
gpt-oss-20b is the stronger model overall, scoring 32.5 to 30.8 on the Noometry Index.
Which is cheaper, Gemma 3 27B or gpt-oss-20b?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Gemma 3 27B lists at $0.08 and $0.16.
Is Gemma 3 27B or gpt-oss-20b better for coding?
gpt-oss-20b scores higher on coding benchmarks: 37.6 versus 22.5 in the Noometry coding category.
Which has the bigger context window?
Both accept 131K tokens.
How many benchmarks do Gemma 3 27B and gpt-oss-20b share?
26 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and gpt-oss-20b has 34.