Model comparison
DeepSeek-R1 vs Gemma 3 4B
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 28.1 on the Noometry Index. Gemma 3 4B costs 18× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Last verified . 18 shared benchmarks.
DeepSeek
GoogleSummary
- They share 18 benchmarks with published results for both. DeepSeek-R1 scores higher in 9 categories and Gemma 3 4B in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-R1 leads 44.5 to 11.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 66.4% for DeepSeek-R1 and 7.5% for Gemma 3 4B.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- DeepSeek-R1 accepts more context: 164K tokens versus 131K.
- Gemma 3 4B has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | Gemma 3 4B | |
|---|---|---|
| Provider | DeepSeek | |
| Noometry Index | 42.3 | 28.1 |
| Released | 2025-01-20 | 2025-03-12 |
| Weights | Proprietary | Open |
| Context window | 164K | 131K |
| Max output | 64K | 4K |
| Input $ / M tokens | $0.50 | $0.04 |
| Output $ / M tokens | $2.15 | $0.08 |
| Results tracked | 52 | 22 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), Gemma 3 4B: 35.9 (#215)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| LMArena Coding | 1427 | 1230 |
| Aider Polyglot | 71.4% | — |
| SciCode | 35.7% | — |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| ALE-Bench | 804.12 | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.7 (#75), Gemma 3 4B: 20.9 (#142)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 19.6% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| METR Time Horizons | 53.8% | — |
Reasoning DeepSeek-R1 leads
DeepSeek-R1: 18.6 (#278), Gemma 3 4B: 13.2 (#335)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| Kagi LLM Benchmark | 69.4% | 25.2% |
| LMArena Hard Prompts | 1416 | 1253 |
| Epoch Capabilities Index | 141.29 | 116.02 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| ARC-AGI-1 | 21.2% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | — | 0% |
| LiveBench Reasoning | 83.2% | — |
| DTBench | — | 50.9% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 2.8% |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), Gemma 3 4B: 16.8 (#292)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 7.5% |
| LMArena Math | 1400 | 1239 |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.5 (#87), Gemma 3 4B: 11.8 (#299)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| GPQA Diamond | 76.3% | 23.2% |
| Vectara Hallucination Rate | 11.3% | 6.4% |
| LMArena Expert | 1394 | 1223 |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| GPQA (HELM) | 66.6% | — |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.4 (#85), Gemma 3 4B: 42.5 (#194)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| LMArena Non-English | 1412 | 1273 |
| LMArena German | 1404 | 1281 |
| LMArena Russian | 1423 | 1294 |
| LMArena Chinese | 1442 | — |
| LMArena French | 1417 | — |
| LMArena Japanese | 1391 | — |
| LMArena Korean | 1360 | — |
| LMArena Spanish | 1411 | — |
Instruction Following DeepSeek-R1 leads
DeepSeek-R1: 72.0 (#143), Gemma 3 4B: 65.2 (#225)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| LMArena Instruction Following | 1382 | 1239 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), Gemma 3 4B: 38.7 (#194)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| LMArena Longer Query | 1391 | 1273 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.4 (#88), Gemma 3 4B: 42.0 (#239)
| Benchmark | DeepSeek-R1 | Gemma 3 4B |
|---|---|---|
| LMArena Text | 1428 | 1291 |
| LMArena Creative Writing | 1405 | 1271 |
| EQ-Bench Creative Writing | 1500 | 1068 |
| LMArena Multi-Turn | 1405 | 1255 |
| Short-Story Creative Writing | 83% | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than Gemma 3 4B?
DeepSeek-R1 is the stronger model overall, scoring 42.3 to 28.1 on the Noometry Index. Gemma 3 4B costs 18× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 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; DeepSeek-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or Gemma 3 4B better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 35.9 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-R1 does, with 164K tokens against 131K.
How many benchmarks do DeepSeek-R1 and Gemma 3 4B share?
18 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Gemma 3 4B has 22.