Model comparison
Grok 3 vs o3-pro
o3-pro is the stronger model overall, scoring 42.9 to 39.9 on the Noometry Index.
Last verified . 10 shared benchmarks.
Summary
- They share 10 benchmarks with published results for both. Grok 3 scores higher in 1 category and o3-pro in 4 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in long context, where o3-pro leads 72.2 to 38.7.
- The biggest single-benchmark swing is ARC-AGI-1: 5.5% for Grok 3 and 59.3% for o3-pro.
Side by side
| Grok 3 | o3-pro | |
|---|---|---|
| Provider | xAI | OpenAI |
| Noometry Index | 39.9 | 42.9 |
| Released | 2025-04-09 | 2025-06-10 |
| Weights | Proprietary | Proprietary |
| Context window | — | 200K |
| Max output | — | 100K |
| Input $ / M tokens | — | $20 |
| Output $ / M tokens | — | $80 |
| Results tracked | 40 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3-pro leads
Grok 3: 41.9 (#115), o3-pro: 55.5 (#24)
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| Aider Polyglot | 53.3% | 84.9% |
| WeirdML | 37.2% | 58.2% |
| LMArena Coding | 1432 | — |
Agentic & Tool Use Not comparable
Grok 3: 30.5 (#76), o3-pro: —
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| BALROG | 29.5% | — |
Reasoning o3-pro leads
Grok 3: 13.7 (#333), o3-pro: 23.8 (#171)
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| ARC-AGI-2 | 0% | 4.9% |
| Kagi LLM Benchmark | 61.3% | 72.1% |
| ARC-AGI-1 | 5.5% | 59.3% |
| Epoch Capabilities Index | 138.33 | 147.42 |
| SimpleBench | 36.1% | — |
| LMArena Hard Prompts | 1434 | — |
| DTBench | — | 86.9% |
| LMCA | — | 38.5% |
Math Not comparable
Grok 3: 38.0 (#145), o3-pro: —
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 55.6% | — |
| Omni-MATH | 46.4% | — |
| LMArena Math | 1391 | — |
| MATH Level 5 | 88.7% | — |
| FrontierMath (Feb 2025 set) | 3.8% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge Grok 3 leads
Grok 3: 46.2 (#82), o3-pro: 29.5 (#238)
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| Confabulations | 14.2% | 14.2% |
| Vectara Hallucination Rate | 5.8% | 23.3% |
| GPQA Diamond | 75.8% | — |
| MMLU-Pro | 78.8% | — |
| GPQA (HELM) | 65% | — |
| LMArena Expert | 1421 | — |
Multilingual Not comparable
Grok 3: 52.3 (#87), o3-pro: —
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| LMArena Non-English | 1410 | — |
| LMArena Chinese | 1448 | — |
| LMArena French | 1460 | — |
| LMArena German | 1431 | — |
| LMArena Japanese | 1387 | — |
| LMArena Korean | 1373 | — |
| LMArena Russian | 1416 | — |
| LMArena Spanish | 1417 | — |
Instruction Following Not comparable
Grok 3: 75.0 (#73), o3-pro: —
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| IFEval | 88.4% | — |
| LMArena Instruction Following | 1409 | — |
Long Context o3-pro leads
Grok 3: 38.7 (#192), o3-pro: 72.2 (#1)
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| Fiction.LiveBench | 58.3% | 97.2% |
| LMArena Longer Query | 1439 | — |
Writing & Preference o3-pro leads
Grok 3: 55.8 (#141), o3-pro: 57.1 (#133)
| Benchmark | Grok 3 | o3-pro |
|---|---|---|
| Short-Story Creative Writing | 76.4% | 84.4% |
| LMArena Text | 1426 | — |
| LMArena Creative Writing | 1414 | — |
| EQ-Bench Creative Writing | 1186 | — |
| WildBench | 84.9% | — |
| LMArena Multi-Turn | 1425 | — |
Frequently asked questions
Is Grok 3 better than o3-pro?
o3-pro is the stronger model overall, scoring 42.9 to 39.9 on the Noometry Index.
Is Grok 3 or o3-pro better for coding?
o3-pro scores higher on coding benchmarks: 55.5 versus 41.9 in the Noometry coding category.
How many benchmarks do Grok 3 and o3-pro share?
10 benchmarks have published results for both models. Grok 3 has 40 scored results on Noometry and o3-pro has 12.