Model comparison
Claude Opus 4 vs Claude Sonnet 4
Claude Opus 4 is the stronger model overall, scoring 43.1 to 40.8 on the Noometry Index. Claude Sonnet 4 costs 5.0× less per token, which makes it the better buy when Claude Opus 4's lead doesn't matter for your workload.
Last verified . 53 shared benchmarks.
Summary
- They share 53 benchmarks with published results for both. Claude Opus 4 scores higher in 8 categories and Claude Sonnet 4 in 2 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in long context, where Claude Opus 4 leads 39.6 to 33.7.
- The biggest single-benchmark swing is Fiction.LiveBench: 61.1% for Claude Opus 4 and 46.9% for Claude Sonnet 4.
- Claude Sonnet 4 is cheaper at $3 / $15 per million input/output tokens, against $15 / $75 for Claude Opus 4.
Side by side
| Claude Opus 4 | Claude Sonnet 4 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 43.1 | 40.8 |
| Released | 2025-05-22 | 2025-05-22 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 200K |
| Max output | 32K | 64K |
| Input $ / M tokens | $15 | $3 |
| Output $ / M tokens | $75 | $15 |
| Results tracked | 56 | 58 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 4 leads
Claude Opus 4: 47.2 (#62), Claude Sonnet 4: 43.5 (#88)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| SWE-bench Verified (bash only) | 67.6% | 64.9% |
| Aider Polyglot | 72% | 61.3% |
| GSO | 6.9% | 4.9% |
| WeirdML | 43.7% | 46.1% |
| LMArena Coding | 1442 | 1414 |
| SWE-bench Verified | 70.7% | — |
| SciCode | — | 40% |
| ALE-Bench | — | 655.35 |
| AlgoTune | 1.33 | — |
Agentic & Tool Use Claude Sonnet 4 leads
Claude Opus 4: 34.8 (#42), Claude Sonnet 4: 38.5 (#31)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| Cybench | 38% | 35% |
| DeepResearch Bench | 46.8% | 46.6% |
| METR Time Horizons | 63.9% | 62% |
| TheAgentCompany | — | 33.1% |
| OSWorld | — | 43.9% |
| LMArena Search | 1127 | — |
Reasoning Claude Opus 4 leads
Claude Opus 4: 27.3 (#121), Claude Sonnet 4: 22.9 (#187)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| ARC-AGI-2 | 8.6% | 5.9% |
| SimpleBench | 58.8% | 45.5% |
| Kagi LLM Benchmark | 74.3% | 73% |
| ARC-AGI-1 | 35.7% | 40% |
| CritPt | 0.3% | 0.3% |
| EnigmaEval | 5.6% | 3.1% |
| LMArena Hard Prompts | 1399 | 1372 |
| DTBench | 81.6% | 77.1% |
| LMCA | 37.4% | 29% |
| Epoch Capabilities Index | 142.67 | 141.69 |
| ForecastBench | 61.1 | 60.2 |
Math Claude Sonnet 4 leads
Claude Opus 4: 42.0 (#86), Claude Sonnet 4: 43.3 (#80)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 64.4% | 71.1% |
| Omni-MATH | 61.6% | 60.2% |
| LMArena Math | 1390 | 1375 |
| MATH Level 5 | 85% | 84.4% |
| FrontierMath (Feb 2025 set) | 4.5% | 4.1% |
| FrontierMath Tier 4 (v1) | 4.2% | 0% |
Knowledge Claude Opus 4 leads
Claude Opus 4: 44.0 (#88), Claude Sonnet 4: 41.8 (#108)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| GPQA Diamond | 76.3% | 79.2% |
| Humanity's Last Exam | 10.7% | 7.8% |
| MMLU-Pro | 87.5% | 84.3% |
| Confabulations | 15.9% | 13.2% |
| Vectara Hallucination Rate | 12% | 10.3% |
| GPQA (HELM) | 70.8% | 70.6% |
| LMArena Expert | 1386 | 1372 |
Multimodal Claude Opus 4 leads
Claude Opus 4: 31.5 (#106), Claude Sonnet 4: 26.2 (#121)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| LMArena Vision | 1192 | 1191 |
| GeoBench | 49% | 37% |
| VPCT | 38% | 34% |
| MindCube | — | 44.8% |
Multilingual Claude Opus 4 leads
Claude Opus 4: 48.8 (#138), Claude Sonnet 4: 46.7 (#156)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| LMArena Non-English | 1362 | 1333 |
| LMArena Chinese | 1386 | 1350 |
| LMArena French | 1372 | 1363 |
| LMArena German | 1391 | 1331 |
| LMArena Japanese | 1331 | 1302 |
| LMArena Korean | 1321 | 1291 |
| LMArena Russian | 1392 | 1355 |
| LMArena Spanish | 1389 | 1357 |
Instruction Following Claude Opus 4 leads
Claude Opus 4: 77.1 (#28), Claude Sonnet 4: 71.7 (#145)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| IFEval | 91.8% | 84% |
| LMArena Instruction Following | 1406 | 1376 |
Long Context Claude Opus 4 leads
Claude Opus 4: 39.6 (#172), Claude Sonnet 4: 33.7 (#259)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| Fiction.LiveBench | 61.1% | 46.9% |
| LMArena Longer Query | 1422 | 1398 |
Writing & Preference Claude Opus 4 leads
Claude Opus 4: 61.2 (#89), Claude Sonnet 4: 57.1 (#132)
| Benchmark | Claude Opus 4 | Claude Sonnet 4 |
|---|---|---|
| LMArena Text | 1377 | 1351 |
| LMArena Creative Writing | 1387 | 1345 |
| Short-Story Creative Writing | 83.6% | 81.4% |
| EQ-Bench Creative Writing | 1580 | 1483 |
| WildBench | 85.2% | 83.8% |
| LMArena Multi-Turn | 1396 | 1376 |
Frequently asked questions
Is Claude Opus 4 better than Claude Sonnet 4?
Claude Opus 4 is the stronger model overall, scoring 43.1 to 40.8 on the Noometry Index. Claude Sonnet 4 costs 5.0× less per token, which makes it the better buy when Claude Opus 4's lead doesn't matter for your workload.
Which is cheaper, Claude Opus 4 or Claude Sonnet 4?
Claude Sonnet 4 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; Claude Opus 4 lists at $15 and $75.
Is Claude Opus 4 or Claude Sonnet 4 better for coding?
Claude Opus 4 scores higher on coding benchmarks: 47.2 versus 43.5 in the Noometry coding category.
Which has the bigger context window?
Both accept 200K tokens.
How many benchmarks do Claude Opus 4 and Claude Sonnet 4 share?
53 benchmarks have published results for both models. Claude Opus 4 has 56 scored results on Noometry and Claude Sonnet 4 has 58.