Model comparison
Gemma 3 4B vs Grok 4.3
Grok 4.3 is the stronger model overall, scoring 43.6 to 27.9 on the Noometry Index. Gemma 3 4B costs 31× less per token, which makes it the better buy when Grok 4.3's lead doesn't matter for your workload.
Last verified . 18 shared benchmarks.
GoogleSummary
- They share 18 benchmarks with published results for both. Gemma 3 4B scores higher in 0 categories and Grok 4.3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok 4.3 leads 52.5 to 11.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 7.5% for Gemma 3 4B and 93.3% for Grok 4.3.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $1.25 / $2.50 for Grok 4.3.
- Grok 4.3 accepts more context: 1M tokens versus 131K.
- Gemma 3 4B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 4B | Grok 4.3 | |
|---|---|---|
| Provider | xAI | |
| Noometry Index | 27.9 | 43.6 |
| Released | 2025-03-12 | 2026-04-17 |
| Weights | Open | Proprietary |
| Context window | 131K | 1M |
| Max output | 4K | 30K |
| Input $ / M tokens | $0.04 | $1.25 |
| Output $ / M tokens | $0.08 | $2.50 |
| Results tracked | 22 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.3 leads
Gemma 3 4B: 35.3 (#214), Grok 4.3: 41.3 (#117)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| LMArena Coding | 1230 | 1415 |
| LMArena WebDev | — | 1357 |
| SciCode | — | 47.3% |
| WeirdML | — | 49.9% |
| ALE-Bench | — | 944.17 |
Agentic & Tool Use Grok 4.3 leads
Gemma 3 4B: 20.7 (#134), Grok 4.3: 27.6 (#93)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | — |
| GDP.pdf | — | 8% |
| LMArena Search | — | 1165 |
| Vending-Bench 2 | — | 35.26 |
Reasoning Grok 4.3 leads
Gemma 3 4B: 12.7 (#330), Grok 4.3: 34.9 (#60)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| Chess Puzzles | 0% | 25% |
| LMArena Hard Prompts | 1253 | 1397 |
| DTBench | 50.9% | 90.7% |
| LMCA | 2.8% | 38.3% |
| Epoch Capabilities Index | 116.02 | 149.13 |
| Kagi LLM Benchmark | 25.2% | — |
| NYT Connections (extended) | — | 55.2% |
| CritPt | — | 8% |
| ForecastBench | — | 60.3 |
Math Grok 4.3 leads
Gemma 3 4B: 15.8 (#286), Grok 4.3: 45.8 (#60)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 93.3% |
| LMArena Math | 1239 | 1388 |
| FrontierMath (Tiers 1-3) | — | 42.8% |
| FrontierMath Tier 4 | — | 14.6% |
| ProofBench | — | 11% |
Knowledge Grok 4.3 leads
Gemma 3 4B: 11.8 (#294), Grok 4.3: 52.5 (#57)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| GPQA Diamond | 23.2% | 88.8% |
| LMArena Expert | 1223 | 1385 |
| SimpleQA Verified | — | 33.2% |
| Vectara Hallucination Rate | 6.4% | — |
Multimodal Not comparable
Gemma 3 4B: —, Grok 4.3: 31.6 (#96)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| LMArena Vision | — | 1231 |
| Blueprint-Bench 2 | — | 0% |
Multilingual Grok 4.3 leads
Gemma 3 4B: 42.7 (#189), Grok 4.3: 50.7 (#114)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| LMArena Non-English | 1273 | 1385 |
| LMArena German | 1281 | 1395 |
| LMArena Russian | 1294 | 1399 |
| LMArena Chinese | — | 1423 |
| LMArena French | — | 1412 |
| LMArena Japanese | — | 1379 |
| LMArena Korean | — | 1357 |
| LMArena Spanish | — | 1402 |
Instruction Following Grok 4.3 leads
Gemma 3 4B: 65.4 (#219), Grok 4.3: 72.3 (#135)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| LMArena Instruction Following | 1239 | 1366 |
Long Context Grok 4.3 leads
Gemma 3 4B: 38.6 (#189), Grok 4.3: 42.5 (#118)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| LMArena Longer Query | 1274 | 1393 |
Writing & Preference Grok 4.3 leads
Gemma 3 4B: 42.2 (#234), Grok 4.3: 58.6 (#112)
| Benchmark | Gemma 3 4B | Grok 4.3 |
|---|---|---|
| LMArena Text | 1291 | 1397 |
| LMArena Creative Writing | 1270 | 1380 |
| LMArena Multi-Turn | 1255 | 1406 |
| EQ-Bench Creative Writing | 1068 | — |
| EQ-Bench 4 | — | 1075 |
Frequently asked questions
Is Gemma 3 4B better than Grok 4.3?
Grok 4.3 is the stronger model overall, scoring 43.6 to 27.9 on the Noometry Index. Gemma 3 4B costs 31× less per token, which makes it the better buy when Grok 4.3's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 4B or Grok 4.3?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; Grok 4.3 lists at $1.25 and $2.50.
Is Gemma 3 4B or Grok 4.3 better for coding?
Grok 4.3 scores higher on coding benchmarks: 41.3 versus 35.3 in the Noometry coding category.
Which has the bigger context window?
Grok 4.3 does, with 1M tokens against 131K.
How many benchmarks do Gemma 3 4B and Grok 4.3 share?
18 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Grok 4.3 has 40.