Model comparison
Grok 4.20 (Non-Reasoning) vs Grok 4.3
Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 43.8 on the Noometry Index.
Last verified . 36 shared benchmarks.
Summary
- They share 36 benchmarks with published results for both. Grok 4.20 (Non-Reasoning) scores higher in 10 categories and Grok 4.3 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.20 (Non-Reasoning) leads 52.3 to 35.9.
- The biggest single-benchmark swing is NYT Connections (extended): 85.4% for Grok 4.20 (Non-Reasoning) and 55.2% for Grok 4.3.
- Both cost about the same: $1.25 input and $2.50 output per million tokens.
Side by side
| Grok 4.20 (Non-Reasoning) | Grok 4.3 | |
|---|---|---|
| Provider | xAI | xAI |
| Noometry Index | 48.6 | 43.8 |
| Released | 2026-02-17 | 2026-04-17 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1M |
| Max output | 30K | 30K |
| Input $ / M tokens | $1.25 | $1.25 |
| Output $ / M tokens | $2.50 | $2.50 |
| Results tracked | 46 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Grok 4.20 (Non-Reasoning): 42.1 (#112), Grok 4.3: 41.6 (#121)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| LMArena WebDev | 1375 | 1357 |
| WeirdML | 52.3% | 49.9% |
| LMArena Coding | 1459 | 1415 |
| ALE-Bench | 1,150 | 944.17 |
| SciCode | — | 47.3% |
Agentic & Tool Use Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 34.4 (#46), Grok 4.3: 27.7 (#99)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| LMArena Search | 1189 | 1165 |
| Vending-Bench 2 | 4,663 | 35.26 |
| Terminal-Bench | 57.3% | — |
| τ²-bench Banking | 18% | — |
| GDP.pdf | — | 8% |
Reasoning Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 52.3 (#32), Grok 4.3: 35.9 (#68)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| NYT Connections (extended) | 85.4% | 55.2% |
| Chess Puzzles | 24% | 25% |
| LMArena Hard Prompts | 1451 | 1396 |
| DTBench | 90.1% | 90.7% |
| LMCA | 38.7% | 38.3% |
| Epoch Capabilities Index | 151.98 | 149.16 |
| ForecastBench | 61.4 | 60.3 |
| ARC-AGI-2 | 65.1% | — |
| Kagi LLM Benchmark | 75% | — |
| ARC-AGI-1 | 89.5% | — |
| CritPt | — | 8% |
| Thematic Generalization | 63.8% | — |
Math Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 48.2 (#65), Grok 4.3: 46.0 (#74)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 44.9% | 42.8% |
| FrontierMath Tier 4 | 17.1% | 14.6% |
| OTIS Mock AIME 2024-2025 | 92.2% | 93.3% |
| ProofBench | 14% | 11% |
| LMArena Math | 1455 | 1388 |
Knowledge Too close to call
Grok 4.20 (Non-Reasoning): 52.8 (#60), Grok 4.3: 52.5 (#62)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| GPQA Diamond | 89.3% | 88.8% |
| SimpleQA Verified | 30.2% | 33.2% |
| LMArena Expert | 1439 | 1385 |
Multimodal Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 33.3 (#98), Grok 4.3: 31.6 (#104)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| LMArena Vision | 1263 | 1229 |
| Blueprint-Bench 2 | 0% | 0% |
| LMArena Document | 1416 | — |
Multilingual Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 54.5 (#40), Grok 4.3: 50.5 (#120)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| LMArena Non-English | 1441 | 1385 |
| LMArena Chinese | 1481 | 1422 |
| LMArena French | 1476 | 1412 |
| LMArena German | 1465 | 1395 |
| LMArena Japanese | 1449 | 1379 |
| LMArena Korean | 1417 | 1356 |
| LMArena Russian | 1458 | 1399 |
| LMArena Spanish | 1443 | 1398 |
Instruction Following Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 74.8 (#83), Grok 4.3: 72.1 (#140)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| LMArena Instruction Following | 1420 | 1366 |
Long Context Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 45.5 (#34), Grok 4.3: 42.5 (#123)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| LMArena Longer Query | 1437 | 1393 |
| 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 4.3: 58.5 (#118)
| Benchmark | Grok 4.20 (Non-Reasoning) | Grok 4.3 |
|---|---|---|
| LMArena Text | 1451 | 1397 |
| LMArena Creative Writing | 1438 | 1380 |
| LMArena Multi-Turn | 1456 | 1406 |
| EQ-Bench Creative Writing | 1574 | — |
| EQ-Bench 4 | — | 1075 |
Frequently asked questions
Is Grok 4.20 (Non-Reasoning) better than Grok 4.3?
Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 43.8 on the Noometry Index.
Which is cheaper, Grok 4.20 (Non-Reasoning) or Grok 4.3?
Grok 4.3 is cheaper. It lists at $1.25 per million input tokens and $2.50 per million output tokens; Grok 4.20 (Non-Reasoning) lists at $1.25 and $2.50.
Is Grok 4.20 (Non-Reasoning) or Grok 4.3 better for coding?
They score almost the same on coding (42.1 vs 41.6); test both on your own repository before choosing.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do Grok 4.20 (Non-Reasoning) and Grok 4.3 share?
36 benchmarks have published results for both models. Grok 4.20 (Non-Reasoning) has 46 scored results on Noometry and Grok 4.3 has 40.