Model comparison
Gemma 3n E4b IT vs Nemotron 4 340b Instruct
Gemma 3n E4b IT is the stronger model overall, scoring 37.3 to 35.9 on the Noometry Index.
Last verified . 17 shared benchmarks.
Google
NVIDIASummary
- They share 17 benchmarks with published results for both. Gemma 3n E4b IT scores higher in 7 categories and Nemotron 4 340b Instruct in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Gemma 3n E4b IT leads 50.1 to 42.7.
Side by side
| Gemma 3n E4b IT | Nemotron 4 340b Instruct | |
|---|---|---|
| Provider | NVIDIA | |
| Noometry Index | 37.3 | 35.9 |
| Released | — | 2024-06-14 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 18 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemma 3n E4b IT leads
Gemma 3n E4b IT: 37.0 (#198), Nemotron 4 340b Instruct: 35.2 (#228)
| Benchmark | Gemma 3n E4b IT | Nemotron 4 340b Instruct |
|---|---|---|
| LMArena Coding | 1268 | 1209 |
Reasoning Nemotron 4 340b Instruct leads
Gemma 3n E4b IT: 19.9 (#247), Nemotron 4 340b Instruct: 23.4 (#178)
| Benchmark | Gemma 3n E4b IT | Nemotron 4 340b Instruct |
|---|---|---|
| LMArena Hard Prompts | 1284 | 1202 |
| Kagi LLM Benchmark | 31.5% | — |
Math Too close to call
Gemma 3n E4b IT: 35.1 (#188), Nemotron 4 340b Instruct: 34.3 (#195)
| Benchmark | Gemma 3n E4b IT | Nemotron 4 340b Instruct |
|---|---|---|
| LMArena Math | 1251 | 1216 |
Knowledge Gemma 3n E4b IT leads
Gemma 3n E4b IT: 34.2 (#198), Nemotron 4 340b Instruct: 32.0 (#216)
| Benchmark | Gemma 3n E4b IT | Nemotron 4 340b Instruct |
|---|---|---|
| LMArena Expert | 1246 | 1171 |
Multilingual Gemma 3n E4b IT leads
Gemma 3n E4b IT: 43.4 (#183), Nemotron 4 340b Instruct: 37.9 (#233)
| Benchmark | Gemma 3n E4b IT | Nemotron 4 340b Instruct |
|---|---|---|
| LMArena Non-English | 1285 | 1206 |
| LMArena Chinese | 1309 | 1214 |
| LMArena French | 1330 | 1197 |
| LMArena German | 1311 | 1192 |
| LMArena Japanese | 1272 | 1128 |
| LMArena Korean | 1259 | 1148 |
| LMArena Russian | 1288 | 1219 |
| LMArena Spanish | 1305 | 1195 |
Instruction Following Gemma 3n E4b IT leads
Gemma 3n E4b IT: 66.1 (#210), Nemotron 4 340b Instruct: 63.1 (#232)
| Benchmark | Gemma 3n E4b IT | Nemotron 4 340b Instruct |
|---|---|---|
| LMArena Instruction Following | 1255 | 1203 |
Long Context Gemma 3n E4b IT leads
Gemma 3n E4b IT: 38.7 (#191), Nemotron 4 340b Instruct: 37.1 (#221)
| Benchmark | Gemma 3n E4b IT | Nemotron 4 340b Instruct |
|---|---|---|
| LMArena Longer Query | 1276 | 1225 |
Writing & Preference Gemma 3n E4b IT leads
Gemma 3n E4b IT: 50.1 (#186), Nemotron 4 340b Instruct: 42.7 (#233)
| Benchmark | Gemma 3n E4b IT | Nemotron 4 340b Instruct |
|---|---|---|
| LMArena Text | 1306 | 1225 |
| LMArena Creative Writing | 1287 | 1203 |
| LMArena Multi-Turn | 1276 | 1214 |
Frequently asked questions
Is Gemma 3n E4b IT better than Nemotron 4 340b Instruct?
Gemma 3n E4b IT is the stronger model overall, scoring 37.3 to 35.9 on the Noometry Index.
Is Gemma 3n E4b IT or Nemotron 4 340b Instruct better for coding?
Gemma 3n E4b IT scores higher on coding benchmarks: 37.0 versus 35.2 in the Noometry coding category.
How many benchmarks do Gemma 3n E4b IT and Nemotron 4 340b Instruct share?
17 benchmarks have published results for both models. Gemma 3n E4b IT has 18 scored results on Noometry and Nemotron 4 340b Instruct has 17.