Model comparison
Gemma 4 31B IT vs Grok 4.6
Grok 4.6 is the stronger model overall, scoring 56.8 to 43.2 on the Noometry Index. Gemma 4 31B IT costs 20× less per token, which makes it the better buy when Grok 4.6's lead doesn't matter for your workload.
Last verified . 29 shared benchmarks.
GoogleSummary
- They share 29 benchmarks with published results for both. Gemma 4 31B IT scores higher in 2 categories and Grok 4.6 in 7 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.6 leads 61.3 to 26.4.
- The biggest single-benchmark swing is SimpleQA Verified: 10.4% for Gemma 4 31B IT and 49.3% for Grok 4.6.
- Gemma 4 31B IT is cheaper at $0.09 / $0.34 per million input/output tokens, against $2 / $6 for Grok 4.6.
- Grok 4.6 accepts more context: 500K tokens versus 262K.
- Gemma 4 31B IT has downloadable open weights; the other is API-only.
Side by side
| Gemma 4 31B IT | Grok 4.6 | |
|---|---|---|
| Provider | xAI | |
| Noometry Index | 43.2 | 56.8 |
| Released | 2026-04-02 | 2026-08-12 |
| Weights | Open | Proprietary |
| Context window | 262K | 500K |
| Max output | 33K | 500K |
| Input $ / M tokens | $0.09 | $2 |
| Output $ / M tokens | $0.34 | $6 |
| Results tracked | 35 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.6 leads
Gemma 4 31B IT: 41.9 (#98), Grok 4.6: 58.0 (#14)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| LMArena WebDev | 1364 | 1620 |
| SciCode | 43.4% | 56.5% |
| WeirdML | 52.3% | 67.3% |
| LMArena Coding | 1459 | 1464 |
| ALE-Bench | 925.5 | 1,508 |
| DeepSWE | — | 67.5% |
| FrontierCode | — | 48% |
| CursorBench | — | 41.4% |
| FrontierSWE | — | 25.3% |
Agentic & Tool Use Not comparable
Gemma 4 31B IT: —, Grok 4.6: 39.4 (#25)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| APEX-Agents | — | 65.3% |
| GDP.pdf | — | 17.2% |
| Vending-Bench 2 | — | 9,047 |
Reasoning Grok 4.6 leads
Gemma 4 31B IT: 26.4 (#119), Grok 4.6: 61.3 (#18)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| NYT Connections (extended) | 70.6% | 80% |
| CritPt | 1.4% | 19.7% |
| Chess Puzzles | 5% | 40% |
| LMArena Hard Prompts | 1448 | 1445 |
| DTBench | 82.7% | 97.3% |
| LMCA | 39.3% | 48.5% |
| Epoch Capabilities Index | 142.74 | 156.48 |
| ARC-AGI-2 | — | 67.1% |
| SimpleBench | — | 75.9% |
| Kagi LLM Benchmark | 63.5% | — |
| ARC-AGI-1 | — | 87.5% |
| Thematic Generalization | 53% | — |
| EBR-Bench | — | 30.5% |
| Mystery Game Puzzles | — | 34% |
| Surface Evolver Bench | 30.6% | — |
Math Grok 4.6 leads
Gemma 4 31B IT: 41.3 (#72), Grok 4.6: 66.8 (#19)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 73.3% | 99.2% |
| LMArena Math | 1464 | 1416 |
| FrontierMath (Tiers 1-3) | — | 66% |
| FrontierMath Tier 4 | — | 31.7% |
| ProofBench | — | 51% |
Knowledge Grok 4.6 leads
Gemma 4 31B IT: 37.9 (#141), Grok 4.6: 63.3 (#18)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| GPQA Diamond | 75.8% | 94% |
| SimpleQA Verified | 10.4% | 49.3% |
| LMArena Expert | 1465 | 1463 |
| Vectara Hallucination Rate | 7.4% | — |
Multimodal Grok 4.6 leads
Gemma 4 31B IT: 41.0 (#32), Grok 4.6: 43.9 (#18)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| LMArena Vision | 1278 | 1265 |
| LMArena Document | 1425 | 1452 |
| Blueprint-Bench 2 | — | 33.2% |
| Furniture Assembly | — | 40% |
Multilingual Too close to call
Gemma 4 31B IT: 54.0 (#52), Grok 4.6: 53.3 (#67)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| LMArena Non-English | 1431 | 1422 |
| LMArena Chinese | 1478 | 1479 |
| LMArena French | 1436 | 1458 |
| LMArena Russian | 1462 | 1427 |
| LMArena Spanish | 1443 | 1408 |
| LMArena German | — | 1425 |
| LMArena Japanese | — | 1383 |
| LMArena Korean | — | 1395 |
Instruction Following Too close to call
Gemma 4 31B IT: 75.6 (#58), Grok 4.6: 75.6 (#59)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| LMArena Instruction Following | 1433 | 1431 |
Long Context Too close to call
Gemma 4 31B IT: 44.2 (#67), Grok 4.6: 44.4 (#62)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| LMArena Longer Query | 1446 | 1453 |
Writing & Preference Grok 4.6 leads
Gemma 4 31B IT: 60.7 (#92), Grok 4.6: 62.5 (#77)
| Benchmark | Gemma 4 31B IT | Grok 4.6 |
|---|---|---|
| LMArena Text | 1443 | 1427 |
| LMArena Creative Writing | 1416 | 1430 |
| LMArena Multi-Turn | 1452 | 1422 |
| EQ-Bench Creative Writing | 1368 | — |
| EQ-Bench 4 | 1120 | — |
Frequently asked questions
Is Gemma 4 31B IT better than Grok 4.6?
Grok 4.6 is the stronger model overall, scoring 56.8 to 43.2 on the Noometry Index. Gemma 4 31B IT costs 20× 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 4 31B IT or Grok 4.6?
Gemma 4 31B IT is cheaper. It lists at $0.09 per million input tokens and $0.34 per million output tokens; Grok 4.6 lists at $2 and $6.
Is Gemma 4 31B IT or Grok 4.6 better for coding?
Grok 4.6 scores higher on coding benchmarks: 58.0 versus 41.9 in the Noometry coding category.
Which has the bigger context window?
Grok 4.6 does, with 500K tokens against 262K.
How many benchmarks do Gemma 4 31B IT and Grok 4.6 share?
29 benchmarks have published results for both models. Gemma 4 31B IT has 35 scored results on Noometry and Grok 4.6 has 49.