Model comparison
GPT-4.5 vs o3
o3 is the stronger model overall, scoring 47.5 to 37.2 on the Noometry Index.
Last verified . 34 shared benchmarks.
Summary
- They share 34 benchmarks with published results for both. GPT-4.5 scores higher in 1 category and o3 in 9 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where o3 leads 54.6 to 32.5.
- The biggest single-benchmark swing is ARC-AGI-1: 10.3% for GPT-4.5 and 60.8% for o3.
Side by side
| GPT-4.5 | o3 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 37.2 | 47.5 |
| Released | 2025-02-27 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | — | 200K |
| Max output | — | 100K |
| Input $ / M tokens | — | $2 |
| Output $ / M tokens | — | $8 |
| Results tracked | 42 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
GPT-4.5: 42.2 (#109), o3: 46.8 (#64)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| Aider Polyglot | 44.9% | 81.3% |
| WeirdML | 39.4% | 52.4% |
| LMArena Coding | 1396 | 1408 |
| SWE-bench Verified | — | 62.3% |
| SWE-bench Verified (bash only) | — | 58.4% |
| GSO | — | 8.8% |
| LiveBench Coding | 75.2% | — |
| CadEval | — | 74% |
| ALE-Bench | — | 933.55 |
Agentic & Tool Use o3 leads
GPT-4.5: 27.9 (#97), o3: 34.5 (#44)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| Cybench | 17.5% | — |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
Reasoning o3 leads
GPT-4.5: 13.9 (#330), o3: 32.0 (#78)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| ARC-AGI-2 | 0.8% | 6.5% |
| SimpleBench | 34.5% | 53.1% |
| ARC-AGI-1 | 10.3% | 60.8% |
| EnigmaEval | 3.2% | 13.1% |
| LMArena Hard Prompts | 1403 | 1402 |
| Epoch Capabilities Index | 136.74 | 146.86 |
| ForecastBench | 61.7 | 62.5 |
| Kagi LLM Benchmark | — | 67.6% |
| CritPt | — | 1.4% |
| Chess Puzzles | — | 38% |
| LiveBench Reasoning | 71.1% | — |
| Mystery Game Puzzles | — | 29% |
| DTBench | — | 84.8% |
| LiveBench Data Analysis | 64.3% | — |
| LMCA | — | 39.7% |
| LiveBench | 69% | — |
Math o3 leads
GPT-4.5: 32.6 (#211), o3: 50.2 (#58)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 84.4% |
| LMArena Math | 1412 | 1426 |
| MATH Level 5 | 78.6% | 97.8% |
| FrontierMath (Tiers 1-3) | — | 33.3% |
| Omni-MATH | — | 71.4% |
| LiveBench Math | 69.3% | — |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge o3 leads
GPT-4.5: 32.5 (#211), o3: 54.6 (#52)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| GPQA Diamond | 68.7% | 81.8% |
| Humanity's Last Exam | 5.4% | 20.3% |
| Confabulations | 13.6% | 14.4% |
| LMArena Expert | 1394 | 1402 |
| SimpleQA Verified | — | 49.4% |
| MMLU-Pro | — | 85.9% |
| GPQA (HELM) | — | 75.3% |
Multimodal o3 leads
GPT-4.5: 37.6 (#71), o3: 41.4 (#36)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| LMArena Vision | 1195 | 1214 |
| VPCT | 45% | 52% |
| GeoBench | — | 74% |
Multilingual Too close to call
GPT-4.5: 52.5 (#83), o3: 51.7 (#105)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| LMArena Non-English | 1413 | 1401 |
| LMArena Chinese | 1421 | 1437 |
| LMArena French | 1418 | 1430 |
| LMArena German | 1457 | 1420 |
| LMArena Japanese | 1416 | 1403 |
| LMArena Korean | 1392 | 1370 |
| LMArena Russian | 1419 | 1406 |
| LMArena Spanish | — | 1395 |
Instruction Following Too close to call
GPT-4.5: 72.6 (#134), o3: 72.8 (#127)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| LMArena Instruction Following | 1404 | 1368 |
| LiveBench Instruction Following | 72.3% | — |
| IFEval | — | 86.9% |
Long Context o3 leads
GPT-4.5: 40.4 (#155), o3: 53.3 (#6)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| Fiction.LiveBench | 63.9% | 88.9% |
| LMArena Longer Query | 1406 | 1372 |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
GPT-4.5: 56.9 (#134), o3: 63.5 (#64)
| Benchmark | GPT-4.5 | o3 |
|---|---|---|
| LMArena Text | 1417 | 1410 |
| LMArena Creative Writing | 1394 | 1359 |
| Short-Story Creative Writing | 75.6% | 83.9% |
| EQ-Bench Creative Writing | 1258 | 1676 |
| LMArena Multi-Turn | 1444 | 1405 |
| WildBench | — | 86.1% |
| LiveBench Language | 61.5% | — |
Frequently asked questions
Is GPT-4.5 better than o3?
o3 is the stronger model overall, scoring 47.5 to 37.2 on the Noometry Index.
Is GPT-4.5 or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 42.2 in the Noometry coding category.
How many benchmarks do GPT-4.5 and o3 share?
34 benchmarks have published results for both models. GPT-4.5 has 42 scored results on Noometry and o3 has 63.