Model comparison
Grok 4.20 (Non-Reasoning) vs o3
Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 47.5 on the Noometry Index.
Last verified . 35 shared benchmarks.
Summary
- They share 35 benchmarks with published results for both. Grok 4.20 (Non-Reasoning) scores higher in 4 categories and o3 in 6 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Grok 4.20 (Non-Reasoning) leads 52.3 to 32.0.
- The biggest single-benchmark swing is ARC-AGI-2: 65.1% for Grok 4.20 (Non-Reasoning) and 6.5% for o3.
- Grok 4.20 (Non-Reasoning) is cheaper at $1.25 / $2.50 per million input/output tokens, against $2 / $8 for o3.
- Grok 4.20 (Non-Reasoning) accepts more context: 1M tokens versus 200K.
Side by side
| Grok 4.20 (Non-Reasoning) | o3 | |
|---|---|---|
| Provider | xAI | OpenAI |
| Noometry Index | 48.6 | 47.5 |
| Released | 2026-02-17 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 200K |
| Max output | 30K | 100K |
| Input $ / M tokens | $1.25 | $2 |
| Output $ / M tokens | $2.50 | $8 |
| Results tracked | 46 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
Grok 4.20 (Non-Reasoning): 42.1 (#112), o3: 46.8 (#64)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| WeirdML | 52.3% | 52.4% |
| LMArena Coding | 1459 | 1408 |
| ALE-Bench | 1,150 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| SWE-bench Verified (bash only) | — | 58.4% |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1375 | — |
| GSO | — | 8.8% |
| CadEval | — | 74% |
Agentic & Tool Use Too close to call
Grok 4.20 (Non-Reasoning): 34.4 (#46), o3: 34.5 (#44)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| LMArena Search | 1189 | 1144 |
| Terminal-Bench | 57.3% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| τ²-bench Banking | 18% | — |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | 4,663 | — |
Reasoning Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 52.3 (#32), o3: 32.0 (#78)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| ARC-AGI-2 | 65.1% | 6.5% |
| Kagi LLM Benchmark | 75% | 67.6% |
| ARC-AGI-1 | 89.5% | 60.8% |
| Chess Puzzles | 24% | 38% |
| LMArena Hard Prompts | 1451 | 1402 |
| DTBench | 90.1% | 84.8% |
| LMCA | 38.7% | 39.7% |
| Epoch Capabilities Index | 151.98 | 146.86 |
| ForecastBench | 61.4 | 62.5 |
| SimpleBench | — | 53.1% |
| NYT Connections (extended) | 85.4% | — |
| CritPt | — | 1.4% |
| EnigmaEval | — | 13.1% |
| Thematic Generalization | 63.8% | — |
| Mystery Game Puzzles | — | 29% |
Math o3 leads
Grok 4.20 (Non-Reasoning): 48.2 (#65), o3: 50.2 (#58)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 44.9% | 33.3% |
| OTIS Mock AIME 2024-2025 | 92.2% | 84.4% |
| LMArena Math | 1455 | 1426 |
| FrontierMath Tier 4 | 17.1% | — |
| ProofBench | 14% | — |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge o3 leads
Grok 4.20 (Non-Reasoning): 52.8 (#60), o3: 54.6 (#52)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| GPQA Diamond | 89.3% | 81.8% |
| SimpleQA Verified | 30.2% | 49.4% |
| LMArena Expert | 1439 | 1402 |
| Humanity's Last Exam | — | 20.3% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |
Multimodal o3 leads
Grok 4.20 (Non-Reasoning): 33.3 (#98), o3: 41.4 (#36)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| LMArena Vision | 1263 | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
| Blueprint-Bench 2 | 0% | — |
| LMArena Document | 1416 | — |
Multilingual Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 54.5 (#40), o3: 51.7 (#105)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| LMArena Non-English | 1441 | 1401 |
| LMArena Chinese | 1481 | 1437 |
| LMArena French | 1476 | 1430 |
| LMArena German | 1465 | 1420 |
| LMArena Japanese | 1449 | 1403 |
| LMArena Korean | 1417 | 1370 |
| LMArena Russian | 1458 | 1406 |
| LMArena Spanish | 1443 | 1395 |
Instruction Following Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 74.8 (#83), o3: 72.8 (#127)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| LMArena Instruction Following | 1420 | 1368 |
| IFEval | — | 86.9% |
Long Context o3 leads
Grok 4.20 (Non-Reasoning): 45.5 (#34), o3: 53.3 (#6)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| CL-bench | 22.2% | 17.8% |
| LMArena Longer Query | 1437 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench Life | 11.9% | — |
Writing & Preference Grok 4.20 (Non-Reasoning) leads
Grok 4.20 (Non-Reasoning): 65.7 (#44), o3: 63.5 (#64)
| Benchmark | Grok 4.20 (Non-Reasoning) | o3 |
|---|---|---|
| LMArena Text | 1451 | 1410 |
| LMArena Creative Writing | 1438 | 1359 |
| EQ-Bench Creative Writing | 1574 | 1676 |
| LMArena Multi-Turn | 1456 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| WildBench | — | 86.1% |
Frequently asked questions
Is Grok 4.20 (Non-Reasoning) better than o3?
Grok 4.20 (Non-Reasoning) is the stronger model overall, scoring 48.6 to 47.5 on the Noometry Index.
Which is cheaper, Grok 4.20 (Non-Reasoning) or o3?
Grok 4.20 (Non-Reasoning) is cheaper. It lists at $1.25 per million input tokens and $2.50 per million output tokens; o3 lists at $2 and $8.
Is Grok 4.20 (Non-Reasoning) or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 42.1 in the Noometry coding category.
Which has the bigger context window?
Grok 4.20 (Non-Reasoning) does, with 1M tokens against 200K.
How many benchmarks do Grok 4.20 (Non-Reasoning) and o3 share?
35 benchmarks have published results for both models. Grok 4.20 (Non-Reasoning) has 46 scored results on Noometry and o3 has 63.