Model comparison
Grok 4.3 vs o3
o3 is the stronger model overall, scoring 47.5 to 43.8 on the Noometry Index. Grok 4.3 costs 2.2× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. Grok 4.3 scores higher in 1 category and o3 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where o3 leads 53.3 to 42.5.
- The biggest single-benchmark swing is SimpleQA Verified: 33.2% for Grok 4.3 and 49.4% for o3.
- Grok 4.3 is cheaper at $1.25 / $2.50 per million input/output tokens, against $2 / $8 for o3.
- Grok 4.3 accepts more context: 1M tokens versus 200K.
Side by side
| Grok 4.3 | o3 | |
|---|---|---|
| Provider | xAI | OpenAI |
| Noometry Index | 43.8 | 47.5 |
| Released | 2026-04-17 | 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 | 40 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
Grok 4.3: 41.6 (#121), o3: 46.8 (#64)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| WeirdML | 49.9% | 52.4% |
| LMArena Coding | 1415 | 1408 |
| ALE-Bench | 944.17 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| SWE-bench Verified (bash only) | — | 58.4% |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1357 | — |
| SciCode | 47.3% | — |
| GSO | — | 8.8% |
| CadEval | — | 74% |
Agentic & Tool Use o3 leads
Grok 4.3: 27.7 (#99), o3: 34.5 (#44)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| LMArena Search | 1165 | 1144 |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| GDP.pdf | 8% | — |
| METR Time Horizons | — | 65.4% |
| Vending-Bench 2 | 35.26 | — |
Reasoning Grok 4.3 leads
Grok 4.3: 35.9 (#68), o3: 32.0 (#78)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| CritPt | 8% | 1.4% |
| Chess Puzzles | 25% | 38% |
| LMArena Hard Prompts | 1396 | 1402 |
| DTBench | 90.7% | 84.8% |
| LMCA | 38.3% | 39.7% |
| Epoch Capabilities Index | 149.16 | 146.86 |
| ForecastBench | 60.3 | 62.5 |
| ARC-AGI-2 | — | 6.5% |
| SimpleBench | — | 53.1% |
| Kagi LLM Benchmark | — | 67.6% |
| NYT Connections (extended) | 55.2% | — |
| ARC-AGI-1 | — | 60.8% |
| EnigmaEval | — | 13.1% |
| Mystery Game Puzzles | — | 29% |
Math o3 leads
Grok 4.3: 46.0 (#74), o3: 50.2 (#58)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 42.8% | 33.3% |
| OTIS Mock AIME 2024-2025 | 93.3% | 84.4% |
| LMArena Math | 1388 | 1426 |
| FrontierMath Tier 4 | 14.6% | — |
| ProofBench | 11% | — |
| 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 4.3: 52.5 (#62), o3: 54.6 (#52)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| GPQA Diamond | 88.8% | 81.8% |
| SimpleQA Verified | 33.2% | 49.4% |
| LMArena Expert | 1385 | 1402 |
| Humanity's Last Exam | — | 20.3% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |
Multimodal o3 leads
Grok 4.3: 31.6 (#104), o3: 41.4 (#36)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| LMArena Vision | 1229 | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
| Blueprint-Bench 2 | 0% | — |
Multilingual o3 leads
Grok 4.3: 50.5 (#120), o3: 51.7 (#105)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| LMArena Non-English | 1385 | 1401 |
| LMArena Chinese | 1422 | 1437 |
| LMArena French | 1412 | 1430 |
| LMArena German | 1395 | 1420 |
| LMArena Japanese | 1379 | 1403 |
| LMArena Korean | 1356 | 1370 |
| LMArena Russian | 1399 | 1406 |
| LMArena Spanish | 1398 | 1395 |
Instruction Following Too close to call
Grok 4.3: 72.1 (#140), o3: 72.8 (#127)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| LMArena Instruction Following | 1366 | 1368 |
| IFEval | — | 86.9% |
Long Context o3 leads
Grok 4.3: 42.5 (#123), o3: 53.3 (#6)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| LMArena Longer Query | 1393 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
Grok 4.3: 58.5 (#118), o3: 63.5 (#64)
| Benchmark | Grok 4.3 | o3 |
|---|---|---|
| LMArena Text | 1397 | 1410 |
| LMArena Creative Writing | 1380 | 1359 |
| LMArena Multi-Turn | 1406 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| EQ-Bench Creative Writing | — | 1676 |
| WildBench | — | 86.1% |
| EQ-Bench 4 | 1075 | — |
Frequently asked questions
Is Grok 4.3 better than o3?
o3 is the stronger model overall, scoring 47.5 to 43.8 on the Noometry Index. Grok 4.3 costs 2.2× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Which is cheaper, Grok 4.3 or o3?
Grok 4.3 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 4.3 or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 41.6 in the Noometry coding category.
Which has the bigger context window?
Grok 4.3 does, with 1M tokens against 200K.
How many benchmarks do Grok 4.3 and o3 share?
31 benchmarks have published results for both models. Grok 4.3 has 40 scored results on Noometry and o3 has 63.