Model comparison
Amazon Nova Pro vs Gemma 3 4B
Amazon Nova Pro is the stronger model overall, scoring 31.0 to 28.1 on the Noometry Index. Gemma 3 4B costs 28× less per token, which makes it the better buy when Amazon Nova Pro's lead doesn't matter for your workload.
Last verified . 15 shared benchmarks.
GoogleSummary
- They share 15 benchmarks with published results for both. Amazon Nova Pro scores higher in 4 categories and Gemma 3 4B in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Amazon Nova Pro leads 27.4 to 11.8.
- The biggest single-benchmark swing is Berkeley Function Calling Leaderboard: 25% for Amazon Nova Pro and 19.6% for Gemma 3 4B.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.80 / $3.20 for Amazon Nova Pro.
- Amazon Nova Pro accepts more context: 300K tokens versus 131K.
- Gemma 3 4B has downloadable open weights; the other is API-only.
Side by side
| Amazon Nova Pro | Gemma 3 4B | |
|---|---|---|
| Provider | Amazon | |
| Noometry Index | 31.0 | 28.1 |
| Released | 2024-12-03 | 2025-03-12 |
| Weights | Proprietary | Open |
| Context window | 300K | 131K |
| Max output | 10K | 4K |
| Input $ / M tokens | $0.80 | $0.04 |
| Output $ / M tokens | $3.20 | $0.08 |
| Results tracked | 38 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Amazon Nova Pro: 35.1 (#229), Gemma 3 4B: 35.9 (#215)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| LMArena Coding | 1270 | 1230 |
| LiveBench Coding | 38.1% | — |
Agentic & Tool Use Gemma 3 4B leads
Amazon Nova Pro: 16.7 (#147), Gemma 3 4B: 20.9 (#142)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | 25% | 19.6% |
| TheAgentCompany | 1.7% | — |
Reasoning Amazon Nova Pro leads
Amazon Nova Pro: 20.0 (#243), Gemma 3 4B: 13.2 (#335)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| LMArena Hard Prompts | 1246 | 1253 |
| Epoch Capabilities Index | 123.8 | 116.02 |
| Kagi LLM Benchmark | — | 25.2% |
| Chess Puzzles | — | 0% |
| LiveBench Reasoning | 32.6% | — |
| DTBench | — | 50.9% |
| LiveBench Data Analysis | 48.3% | — |
| LMCA | — | 2.8% |
| LiveBench | 43.5% | — |
Math Amazon Nova Pro leads
Amazon Nova Pro: 28.5 (#243), Gemma 3 4B: 16.8 (#292)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| LMArena Math | 1252 | 1239 |
| OTIS Mock AIME 2024-2025 | — | 7.5% |
| Omni-MATH | 24.2% | — |
| LiveBench Math | 38% | — |
Knowledge Amazon Nova Pro leads
Amazon Nova Pro: 27.4 (#250), Gemma 3 4B: 11.8 (#299)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| Vectara Hallucination Rate | 5.1% | 6.4% |
| LMArena Expert | 1211 | 1223 |
| GPQA Diamond | — | 23.2% |
| Humanity's Last Exam | 4.4% | — |
| MMLU-Pro | 67.3% | — |
| Confabulations | 30.1% | — |
| GPQA (HELM) | 44.6% | — |
| MMLU | 82% | — |
Multimodal Not comparable
Amazon Nova Pro: 25.0 (#126), Gemma 3 4B: —
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| LMArena Vision | 980 | — |
Multilingual Gemma 3 4B leads
Amazon Nova Pro: 39.7 (#223), Gemma 3 4B: 42.5 (#194)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| LMArena Non-English | 1234 | 1273 |
| LMArena German | 1243 | 1281 |
| LMArena Russian | 1240 | 1294 |
| LMArena Chinese | 1244 | — |
| LMArena French | 1271 | — |
| LMArena Japanese | 1200 | — |
| LMArena Korean | 1203 | — |
| LMArena Spanish | 1182 | — |
Instruction Following Too close to call
Amazon Nova Pro: 64.9 (#226), Gemma 3 4B: 65.2 (#225)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| LMArena Instruction Following | 1235 | 1239 |
| LiveBench Instruction Following | 67.1% | — |
| IFEval | 81.5% | — |
Long Context Too close to call
Amazon Nova Pro: 38.1 (#205), Gemma 3 4B: 38.7 (#194)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| LMArena Longer Query | 1255 | 1273 |
Writing & Preference Amazon Nova Pro leads
Amazon Nova Pro: 43.9 (#226), Gemma 3 4B: 42.0 (#239)
| Benchmark | Amazon Nova Pro | Gemma 3 4B |
|---|---|---|
| LMArena Text | 1259 | 1291 |
| LMArena Creative Writing | 1212 | 1271 |
| LMArena Multi-Turn | 1246 | 1255 |
| Short-Story Creative Writing | 60.5% | — |
| EQ-Bench Creative Writing | — | 1068 |
| WildBench | 77.7% | — |
| LiveBench Language | 37% | — |
Frequently asked questions
Is Amazon Nova Pro better than Gemma 3 4B?
Amazon Nova Pro is the stronger model overall, scoring 31.0 to 28.1 on the Noometry Index. Gemma 3 4B costs 28× less per token, which makes it the better buy when Amazon Nova Pro's lead doesn't matter for your workload.
Which is cheaper, Amazon Nova Pro or Gemma 3 4B?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; Amazon Nova Pro lists at $0.80 and $3.20.
Is Amazon Nova Pro or Gemma 3 4B better for coding?
They score almost the same on coding (35.1 vs 35.9); test both on your own repository before choosing.
Which has the bigger context window?
Amazon Nova Pro does, with 300K tokens against 131K.
How many benchmarks do Amazon Nova Pro and Gemma 3 4B share?
15 benchmarks have published results for both models. Amazon Nova Pro has 38 scored results on Noometry and Gemma 3 4B has 22.