Model comparison
Gemma 3 4B vs Grok 4.6
Grok 4.6 is the stronger model overall, scoring 56.9 to 28.1 on the Noometry Index. Gemma 3 4B costs 60× less per token, which makes it the better buy when Grok 4.6'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 4.6 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok 4.6 leads 63.3 to 11.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 7.5% for Gemma 3 4B and 99.2% for Grok 4.6.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $2 / $6 for Grok 4.6.
- Grok 4.6 accepts more context: 500K tokens versus 131K.
- Gemma 3 4B has downloadable open weights; the other is API-only.
Side by side
| Gemma 3 4B | Grok 4.6 | |
|---|---|---|
| Provider | xAI | |
| Noometry Index | 28.1 | 56.9 |
| Released | 2025-03-12 | 2026-08-12 |
| Weights | Open | Proprietary |
| Context window | 131K | 500K |
| Max output | 4K | 500K |
| Input $ / M tokens | $0.04 | $2 |
| Output $ / M tokens | $0.08 | $6 |
| Results tracked | 22 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.6 leads
Gemma 3 4B: 35.9 (#215), Grok 4.6: 58.5 (#16)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| LMArena Coding | 1230 | 1465 |
| DeepSWE | — | 67.5% |
| FrontierCode | — | 48% |
| CursorBench | — | 41.4% |
| LMArena WebDev | — | 1617 |
| FrontierSWE | — | 25.3% |
| SciCode | — | 56.5% |
| WeirdML | — | 67.3% |
| ALE-Bench | — | 1,508 |
Agentic & Tool Use Grok 4.6 leads
Gemma 3 4B: 20.9 (#142), Grok 4.6: 39.4 (#27)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| APEX-Agents | — | 65.3% |
| Berkeley Function Calling Leaderboard | 19.6% | — |
| GDP.pdf | — | 17.2% |
| Vending-Bench 2 | — | 9,047 |
Reasoning Grok 4.6 leads
Gemma 3 4B: 13.2 (#335), Grok 4.6: 61.4 (#20)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| Chess Puzzles | 0% | 40% |
| LMArena Hard Prompts | 1253 | 1447 |
| DTBench | 50.9% | 97.3% |
| LMCA | 2.8% | 48.5% |
| Epoch Capabilities Index | 116.02 | 156.44 |
| ARC-AGI-2 | — | 67.1% |
| SimpleBench | — | 75.9% |
| Kagi LLM Benchmark | 25.2% | — |
| NYT Connections (extended) | — | 80% |
| ARC-AGI-1 | — | 87.5% |
| CritPt | — | 19.7% |
| EBR-Bench | — | 30.5% |
| Mystery Game Puzzles | — | 34% |
Math Grok 4.6 leads
Gemma 3 4B: 16.8 (#292), Grok 4.6: 67.0 (#24)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 7.5% | 99.2% |
| LMArena Math | 1239 | 1423 |
| FrontierMath (Tiers 1-3) | — | 66% |
| FrontierMath Tier 4 | — | 31.7% |
| ProofBench | — | 51% |
Knowledge Grok 4.6 leads
Gemma 3 4B: 11.8 (#299), Grok 4.6: 63.3 (#20)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| GPQA Diamond | 23.2% | 94% |
| LMArena Expert | 1223 | 1467 |
| SimpleQA Verified | — | 49.3% |
| Vectara Hallucination Rate | 6.4% | — |
Multimodal Not comparable
Gemma 3 4B: —, Grok 4.6: 43.6 (#23)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| LMArena Vision | — | 1263 |
| Blueprint-Bench 2 | — | 33.2% |
| Furniture Assembly | — | 40% |
| LMArena Document | — | 1452 |
Multilingual Grok 4.6 leads
Gemma 3 4B: 42.5 (#194), Grok 4.6: 53.0 (#74)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| LMArena Non-English | 1273 | 1420 |
| LMArena German | 1281 | 1431 |
| LMArena Russian | 1294 | 1422 |
| LMArena Chinese | — | 1480 |
| LMArena French | — | 1461 |
| LMArena Japanese | — | 1376 |
| LMArena Korean | — | 1397 |
| LMArena Spanish | — | 1404 |
Instruction Following Grok 4.6 leads
Gemma 3 4B: 65.2 (#225), Grok 4.6: 75.4 (#63)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| LMArena Instruction Following | 1239 | 1431 |
Long Context Grok 4.6 leads
Gemma 3 4B: 38.7 (#194), Grok 4.6: 44.5 (#66)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| LMArena Longer Query | 1273 | 1454 |
Writing & Preference Grok 4.6 leads
Gemma 3 4B: 42.0 (#239), Grok 4.6: 62.3 (#80)
| Benchmark | Gemma 3 4B | Grok 4.6 |
|---|---|---|
| LMArena Text | 1291 | 1428 |
| LMArena Creative Writing | 1271 | 1428 |
| LMArena Multi-Turn | 1255 | 1425 |
| EQ-Bench Creative Writing | 1068 | — |
Frequently asked questions
Is Gemma 3 4B better than Grok 4.6?
Grok 4.6 is the stronger model overall, scoring 56.9 to 28.1 on the Noometry Index. Gemma 3 4B costs 60× less per token, which makes it the better buy when Grok 4.6's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 4B or Grok 4.6?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; Grok 4.6 lists at $2 and $6.
Is Gemma 3 4B or Grok 4.6 better for coding?
Grok 4.6 scores higher on coding benchmarks: 58.5 versus 35.9 in the Noometry coding category.
Which has the bigger context window?
Grok 4.6 does, with 500K tokens against 131K.
How many benchmarks do Gemma 3 4B and Grok 4.6 share?
18 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and Grok 4.6 has 49.