Model comparison
Claude Opus 4 vs o3
o3 is the stronger model overall, scoring 47.5 to 43.1 on the Noometry Index.
Last verified . 53 shared benchmarks.
Summary
- They share 53 benchmarks with published results for both. Claude Opus 4 scores higher in 3 categories and o3 in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in long context, where o3 leads 53.3 to 39.6.
- The biggest single-benchmark swing is Fiction.LiveBench: 61.1% for Claude Opus 4 and 88.9% for o3.
- o3 is cheaper at $2 / $8 per million input/output tokens, against $15 / $75 for Claude Opus 4.
Side by side
| Claude Opus 4 | o3 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 43.1 | 47.5 |
| Released | 2025-05-22 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 200K |
| Max output | 32K | 100K |
| Input $ / M tokens | $15 | $2 |
| Output $ / M tokens | $75 | $8 |
| Results tracked | 56 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Claude Opus 4: 47.2 (#62), o3: 46.8 (#64)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| SWE-bench Verified | 70.7% | 62.3% |
| SWE-bench Verified (bash only) | 67.6% | 58.4% |
| Aider Polyglot | 72% | 81.3% |
| GSO | 6.9% | 8.8% |
| WeirdML | 43.7% | 52.4% |
| LMArena Coding | 1442 | 1408 |
| CadEval | — | 74% |
| ALE-Bench | — | 933.55 |
| AlgoTune | 1.33 | — |
Agentic & Tool Use Too close to call
Claude Opus 4: 34.8 (#42), o3: 34.5 (#44)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| DeepResearch Bench | 46.8% | 45.2% |
| LMArena Search | 1127 | 1144 |
| METR Time Horizons | 63.9% | 65.4% |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| Cybench | 38% | — |
| OSWorld | — | 23% |
Reasoning o3 leads
Claude Opus 4: 27.3 (#121), o3: 32.0 (#78)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| ARC-AGI-2 | 8.6% | 6.5% |
| SimpleBench | 58.8% | 53.1% |
| Kagi LLM Benchmark | 74.3% | 67.6% |
| ARC-AGI-1 | 35.7% | 60.8% |
| CritPt | 0.3% | 1.4% |
| EnigmaEval | 5.6% | 13.1% |
| LMArena Hard Prompts | 1399 | 1402 |
| DTBench | 81.6% | 84.8% |
| LMCA | 37.4% | 39.7% |
| Epoch Capabilities Index | 142.67 | 146.86 |
| ForecastBench | 61.1 | 62.5 |
| Chess Puzzles | — | 38% |
| Mystery Game Puzzles | — | 29% |
Math o3 leads
Claude Opus 4: 42.0 (#86), o3: 50.2 (#58)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 64.4% | 84.4% |
| Omni-MATH | 61.6% | 71.4% |
| LMArena Math | 1390 | 1426 |
| MATH Level 5 | 85% | 97.8% |
| FrontierMath (Feb 2025 set) | 4.5% | 18.7% |
| FrontierMath Tier 4 (v1) | 4.2% | 2.1% |
| FrontierMath (Tiers 1-3) | — | 33.3% |
Knowledge o3 leads
Claude Opus 4: 44.0 (#88), o3: 54.6 (#52)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| GPQA Diamond | 76.3% | 81.8% |
| Humanity's Last Exam | 10.7% | 20.3% |
| MMLU-Pro | 87.5% | 85.9% |
| Confabulations | 15.9% | 14.4% |
| GPQA (HELM) | 70.8% | 75.3% |
| LMArena Expert | 1386 | 1402 |
| SimpleQA Verified | — | 49.4% |
| Vectara Hallucination Rate | 12% | — |
Multimodal o3 leads
Claude Opus 4: 31.5 (#106), o3: 41.4 (#36)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| LMArena Vision | 1192 | 1214 |
| GeoBench | 49% | 74% |
| VPCT | 38% | 52% |
Multilingual o3 leads
Claude Opus 4: 48.8 (#138), o3: 51.7 (#105)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| LMArena Non-English | 1362 | 1401 |
| LMArena Chinese | 1386 | 1437 |
| LMArena French | 1372 | 1430 |
| LMArena German | 1391 | 1420 |
| LMArena Japanese | 1331 | 1403 |
| LMArena Korean | 1321 | 1370 |
| LMArena Russian | 1392 | 1406 |
| LMArena Spanish | 1389 | 1395 |
Instruction Following Claude Opus 4 leads
Claude Opus 4: 77.1 (#28), o3: 72.8 (#127)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| IFEval | 91.8% | 86.9% |
| LMArena Instruction Following | 1406 | 1368 |
Long Context o3 leads
Claude Opus 4: 39.6 (#172), o3: 53.3 (#6)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| Fiction.LiveBench | 61.1% | 88.9% |
| LMArena Longer Query | 1422 | 1372 |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
Claude Opus 4: 61.2 (#89), o3: 63.5 (#64)
| Benchmark | Claude Opus 4 | o3 |
|---|---|---|
| LMArena Text | 1377 | 1410 |
| LMArena Creative Writing | 1387 | 1359 |
| Short-Story Creative Writing | 83.6% | 83.9% |
| EQ-Bench Creative Writing | 1580 | 1676 |
| WildBench | 85.2% | 86.1% |
| LMArena Multi-Turn | 1396 | 1405 |
Frequently asked questions
Is Claude Opus 4 better than o3?
o3 is the stronger model overall, scoring 47.5 to 43.1 on the Noometry Index.
Which is cheaper, Claude Opus 4 or o3?
o3 is cheaper. It lists at $2 per million input tokens and $8 per million output tokens; Claude Opus 4 lists at $15 and $75.
Is Claude Opus 4 or o3 better for coding?
They score almost the same on coding (47.2 vs 46.8); test both on your own repository before choosing.
Which has the bigger context window?
Both accept 200K tokens.
How many benchmarks do Claude Opus 4 and o3 share?
53 benchmarks have published results for both models. Claude Opus 4 has 56 scored results on Noometry and o3 has 63.