Model comparison
Grok 4.20 (Non-Reasoning) vs Grok 4.6
Grok 4.6 is the stronger model overall, scoring 56.9 to 48.6 on the Noometry Index. Grok 4.20 (Non-Reasoning) costs 1.9× less per token, which makes it the better buy when Grok 4.6's lead doesn't matter for your workload.
Last verified . 37 shared benchmarks.
Summary
- They share 37 benchmarks with published results for both. Grok 4.20 (Non-Reasoning) scores higher in 3 categories and Grok 4.6 in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Grok 4.6 leads 67.0 to 48.2.
- The biggest single-benchmark swing is ProofBench: 14% for Grok 4.20 (Non-Reasoning) and 51% for Grok 4.6.
- Grok 4.20 (Non-Reasoning) is cheaper at $1.25 / $2.50 per million input/output tokens, against $2 / $6 for Grok 4.6.
- Grok 4.20 (Non-Reasoning) accepts more context: 1M tokens versus 500K.
Side by side
| Grok 4.20 (Non-Reasoning) | Grok 4.6 | |
|---|---|---|
| Provider | xAI | xAI |
| Noometry Index | 48.6 | 56.9 |
| Released | 2026-02-17 | 2026-08-12 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 500K |
| Max output | 30K | 500K |
| Input $ / M tokens | $1.25 | $2 |
| Output $ / M tokens | $2.50 | $6 |
| Results tracked | 46 | 49 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4.6 leads
Grok 4.20 (Non-Reasoning): 42.1 (#112), Grok 4.6: 58.5 (#16)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| LMArena WebDev | 1375 | 1617 |
| WeirdML | 52.3% | 67.3% |
| LMArena Coding | 1459 | 1465 |
| ALE-Bench | 1,150 | 1,508 |
| DeepSWE | — | 67.5% |
| FrontierCode | — | 48% |
| CursorBench | — | 41.4% |
| FrontierSWE | — | 25.3% |
| SciCode | — | 56.5% |
Agentic & Tool Use Grok 4.6 leads
Grok 4.20 (Non-Reasoning): 34.4 (#46), Grok 4.6: 39.4 (#27)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| Vending-Bench 2 | 4,663 | 9,047 |
| Terminal-Bench | 57.3% | — |
| APEX-Agents | — | 65.3% |
| τ²-bench Banking | 18% | — |
| GDP.pdf | — | 17.2% |
| LMArena Search | 1189 | — |
Reasoning Grok 4.6 leads
Grok 4.20 (Non-Reasoning): 52.3 (#32), Grok 4.6: 61.4 (#20)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| ARC-AGI-2 | 65.1% | 67.1% |
| NYT Connections (extended) | 85.4% | 80% |
| ARC-AGI-1 | 89.5% | 87.5% |
| Chess Puzzles | 24% | 40% |
| LMArena Hard Prompts | 1451 | 1447 |
| DTBench | 90.1% | 97.3% |
| LMCA | 38.7% | 48.5% |
| Epoch Capabilities Index | 151.98 | 156.44 |
| SimpleBench | — | 75.9% |
| Kagi LLM Benchmark | 75% | — |
| CritPt | — | 19.7% |
| Thematic Generalization | 63.8% | — |
| EBR-Bench | — | 30.5% |
| Mystery Game Puzzles | — | 34% |
| ForecastBench | 61.4 | — |
Math Grok 4.6 leads
Grok 4.20 (Non-Reasoning): 48.2 (#65), Grok 4.6: 67.0 (#24)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 44.9% | 66% |
| FrontierMath Tier 4 | 17.1% | 31.7% |
| OTIS Mock AIME 2024-2025 | 92.2% | 99.2% |
| ProofBench | 14% | 51% |
| LMArena Math | 1455 | 1423 |
Knowledge Grok 4.6 leads
Grok 4.20 (Non-Reasoning): 52.8 (#60), Grok 4.6: 63.3 (#20)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| GPQA Diamond | 89.3% | 94% |
| SimpleQA Verified | 30.2% | 49.3% |
| LMArena Expert | 1439 | 1467 |
Multimodal Grok 4.6 leads
Grok 4.20 (Non-Reasoning): 33.3 (#98), Grok 4.6: 43.6 (#23)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| LMArena Vision | 1263 | 1263 |
| Blueprint-Bench 2 | 0% | 33.2% |
| LMArena Document | 1416 | 1452 |
| Furniture Assembly | — | 40% |
Multilingual Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 54.5 (#40), Grok 4.6: 53.0 (#74)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| LMArena Non-English | 1441 | 1420 |
| LMArena Chinese | 1481 | 1480 |
| LMArena French | 1476 | 1461 |
| LMArena German | 1465 | 1431 |
| LMArena Japanese | 1449 | 1376 |
| LMArena Korean | 1417 | 1397 |
| LMArena Russian | 1458 | 1422 |
| LMArena Spanish | 1443 | 1404 |
Instruction Following Too close to call
Grok 4.20 (Non-Reasoning): 74.8 (#83), Grok 4.6: 75.4 (#63)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| LMArena Instruction Following | 1420 | 1431 |
Long Context Too close to call
Grok 4.20 (Non-Reasoning): 45.5 (#34), Grok 4.6: 44.5 (#66)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| LMArena Longer Query | 1437 | 1454 |
| CL-bench | 22.2% | — |
| CL-bench Life | 11.9% | — |
Writing & Preference Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 65.7 (#44), Grok 4.6: 62.3 (#80)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.6 |
|---|---|---|
| LMArena Text | 1451 | 1428 |
| LMArena Creative Writing | 1438 | 1428 |
| LMArena Multi-Turn | 1456 | 1425 |
| EQ-Bench Creative Writing | 1574 | — |
Frequently asked questions
Is Grok 4.20 (Non-Reasoning) better than Grok 4.6?
Grok 4.6 is the stronger model overall, scoring 56.9 to 48.6 on the Noometry Index. Grok 4.20 (Non-Reasoning) costs 1.9× less per token, which makes it the better buy when Grok 4.6's lead doesn't matter for your workload.
Which is cheaper, Grok 4.20 (Non-Reasoning) or Grok 4.6?
Grok 4.20 (Non-Reasoning) is cheaper. It lists at $1.25 per million input tokens and $2.50 per million output tokens; Grok 4.6 lists at $2 and $6.
Is Grok 4.20 (Non-Reasoning) or Grok 4.6 better for coding?
Grok 4.6 scores higher on coding benchmarks: 58.5 versus 42.1 in the Noometry coding category.
Which has the bigger context window?
Grok 4.20 (Non-Reasoning) does, with 1M tokens against 500K.
How many benchmarks do Grok 4.20 (Non-Reasoning) and Grok 4.6 share?
37 benchmarks have published results for both models. Grok 4.20 (Non-Reasoning) has 46 scored results on Noometry and Grok 4.6 has 49.