Model comparison
Gemma 2 9B vs GPT-5.4 Pro
GPT-5.4 Pro is the stronger model overall, scoring 58.8 to 25.7 on the Noometry Index.
Last verified . 2 shared benchmarks.
GoogleSummary
- They share 2 benchmarks with published results for both. Gemma 2 9B scores higher in 0 categories and GPT-5.4 Pro in 4 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 Pro leads 72.0 to 9.3.
- The biggest single-benchmark swing is GPQA Diamond: 27.5% for Gemma 2 9B and 94.6% for GPT-5.4 Pro.
- Gemma 2 9B has downloadable open weights; the other is API-only.
Side by side
| Gemma 2 9B | GPT-5.4 Pro | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 25.7 | 58.8 |
| Released | 2024-06-24 | 2026-03-05 |
| Weights | Open | Proprietary |
| Context window | — | 1.05M |
| Max output | — | 128K |
| Input $ / M tokens | — | $30 |
| Output $ / M tokens | — | $180 |
| Results tracked | 35 | 16 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 Pro leads
Gemma 2 9B: 29.0 (#299), GPT-5.4 Pro: 42.6 (#89)
| Benchmark | Gemma 2 9B | GPT-5.4 Pro |
|---|---|---|
| WeirdML | — | 57.4% |
| BigCodeBench Instruct | 34.7% | — |
| LiveBench Coding | 22.5% | — |
| LMArena Coding | 1174 | — |
| BigCodeBench Complete | 40.6% | — |
Reasoning GPT-5.4 Pro leads
Gemma 2 9B: 15.3 (#306), GPT-5.4 Pro: 71.0 (#10)
| Benchmark | Gemma 2 9B | GPT-5.4 Pro |
|---|---|---|
| Epoch Capabilities Index | 119.78 | 159.12 |
| ARC-AGI-2 | — | 83.3% |
| SimpleBench | — | 74.1% |
| ARC-AGI-1 | — | 94.5% |
| CritPt | — | 30% |
| Chess Puzzles | — | 58.6% |
| EnigmaEval | — | 23.8% |
| LiveBench Reasoning | 15.2% | — |
| LMArena Hard Prompts | 1172 | — |
| LiveBench Data Analysis | 36.4% | — |
| LiveBench | 28.7% | — |
| PIQA | 83.7% | — |
Math GPT-5.4 Pro leads
Gemma 2 9B: 9.3 (#312), GPT-5.4 Pro: 72.0 (#17)
| Benchmark | Gemma 2 9B | GPT-5.4 Pro |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 82.5% |
| FrontierMath Tier 4 | — | 58.5% |
| OTIS Mock AIME 2024-2025 | 0.6% | — |
| LiveBench Math | 19.8% | — |
| LMArena Math | 1183 | — |
| MATH Level 5 | 21% | — |
| FrontierMath (Feb 2025 set) | — | 50% |
| FrontierMath Tier 4 (v1) | — | 37.5% |
| GSM8K | 84.9% | — |
Knowledge GPT-5.4 Pro leads
Gemma 2 9B: 9.7 (#300), GPT-5.4 Pro: 68.3 (#6)
| Benchmark | Gemma 2 9B | GPT-5.4 Pro |
|---|---|---|
| GPQA Diamond | 27.5% | 94.6% |
| Humanity's Last Exam | — | 44.3% |
| SimpleQA Verified | — | 46.3% |
| Vectara Hallucination Rate | — | 8.3% |
| LMArena Expert | 1147 | — |
| BoolQ | 85.7% | — |
| MMLU | 72.1% | — |
Multilingual Not comparable
Gemma 2 9B: 36.8 (#233), GPT-5.4 Pro: —
| Benchmark | Gemma 2 9B | GPT-5.4 Pro |
|---|---|---|
| LMArena Non-English | 1188 | — |
| LMArena Chinese | 1185 | — |
| LMArena French | 1189 | — |
| LMArena German | 1186 | — |
| LMArena Japanese | 1144 | — |
| LMArena Korean | 1137 | — |
| LMArena Russian | 1200 | — |
| LMArena Spanish | 1200 | — |
Instruction Following Not comparable
Gemma 2 9B: 57.8 (#264), GPT-5.4 Pro: —
| Benchmark | Gemma 2 9B | GPT-5.4 Pro |
|---|---|---|
| LiveBench Instruction Following | 52.6% | — |
| LMArena Instruction Following | 1178 | — |
Long Context Not comparable
Gemma 2 9B: 36.3 (#228), GPT-5.4 Pro: —
| Benchmark | Gemma 2 9B | GPT-5.4 Pro |
|---|---|---|
| LMArena Longer Query | 1197 | — |
Writing & Preference Not comparable
Gemma 2 9B: 32.3 (#276), GPT-5.4 Pro: —
| Benchmark | Gemma 2 9B | GPT-5.4 Pro |
|---|---|---|
| LMArena Text | 1207 | — |
| LMArena Creative Writing | 1206 | — |
| EQ-Bench Creative Writing | 841 | — |
| LMArena Multi-Turn | 1193 | — |
| LiveBench Language | 25.5% | — |
Frequently asked questions
Is Gemma 2 9B better than GPT-5.4 Pro?
GPT-5.4 Pro is the stronger model overall, scoring 58.8 to 25.7 on the Noometry Index.
Is Gemma 2 9B or GPT-5.4 Pro better for coding?
GPT-5.4 Pro scores higher on coding benchmarks: 42.6 versus 29.0 in the Noometry coding category.
How many benchmarks do Gemma 2 9B and GPT-5.4 Pro share?
2 benchmarks have published results for both models. Gemma 2 9B has 35 scored results on Noometry and GPT-5.4 Pro has 16.