Model comparison
Claude Sonnet 4 vs GPT-4.5
Claude Sonnet 4 is the stronger model overall, scoring 40.6 to 36.9 on the Noometry Index.
Last verified . 35 shared benchmarks.
Summary
- They share 35 benchmarks with published results for both. Claude Sonnet 4 scores higher in 6 categories and GPT-4.5 in 4 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in multimodal, where GPT-4.5 leads 37.0 to 25.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 71.1% for Claude Sonnet 4 and 37.8% for GPT-4.5.
Side by side
| Claude Sonnet 4 | GPT-4.5 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 40.6 | 36.9 |
| Released | 2025-05-22 | 2025-02-27 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | — |
| Max output | 64K | — |
| Input $ / M tokens | $3 | — |
| Output $ / M tokens | $15 | — |
| Results tracked | 58 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4 leads
Claude Sonnet 4: 44.0 (#70), GPT-4.5: 41.8 (#102)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| Aider Polyglot | 61.3% | 44.9% |
| WeirdML | 46.1% | 39.4% |
| LMArena Coding | 1414 | 1397 |
| SWE-bench Verified (bash only) | 64.9% | — |
| SciCode | 40% | — |
| GSO | 4.9% | — |
| LiveBench Coding | — | 75.2% |
| ALE-Bench | 655.35 | — |
Agentic & Tool Use Claude Sonnet 4 leads
Claude Sonnet 4: 38.3 (#29), GPT-4.5: 27.6 (#92)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| Cybench | 35% | 17.5% |
| TheAgentCompany | 33.1% | — |
| DeepResearch Bench | 46.6% | — |
| OSWorld | 43.9% | — |
| METR Time Horizons | 62% | — |
Reasoning Claude Sonnet 4 leads
Claude Sonnet 4: 23.0 (#169), GPT-4.5: 14.1 (#320)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| ARC-AGI-2 | 5.9% | 0.8% |
| SimpleBench | 45.5% | 34.5% |
| ARC-AGI-1 | 40% | 10.3% |
| EnigmaEval | 3.1% | 3.2% |
| LMArena Hard Prompts | 1372 | 1404 |
| Epoch Capabilities Index | 141.69 | 136.74 |
| ForecastBench | 60.2 | 61.7 |
| Kagi LLM Benchmark | 73% | — |
| CritPt | 0.3% | — |
| LiveBench Reasoning | — | 71.1% |
| DTBench | 77.1% | — |
| LiveBench Data Analysis | — | 64.3% |
| LMCA | 29% | — |
| LiveBench | — | 69% |
Math Claude Sonnet 4 leads
Claude Sonnet 4: 41.5 (#71), GPT-4.5: 31.0 (#200)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 37.8% |
| LMArena Math | 1375 | 1412 |
| MATH Level 5 | 84.4% | 78.6% |
| Omni-MATH | 60.2% | — |
| LiveBench Math | — | 69.3% |
| FrontierMath (Feb 2025 set) | 4.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge Claude Sonnet 4 leads
Claude Sonnet 4: 41.6 (#99), GPT-4.5: 32.4 (#205)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| GPQA Diamond | 79.2% | 68.7% |
| Humanity's Last Exam | 7.8% | 5.4% |
| Confabulations | 13.2% | 13.6% |
| LMArena Expert | 1371 | 1394 |
| MMLU-Pro | 84.3% | — |
| Vectara Hallucination Rate | 10.3% | — |
| GPQA (HELM) | 70.6% | — |
Multimodal GPT-4.5 leads
Claude Sonnet 4: 25.6 (#116), GPT-4.5: 37.0 (#68)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| LMArena Vision | 1191 | 1195 |
| VPCT | 34% | 45% |
| GeoBench | 37% | — |
| MindCube | 44.8% | — |
Multilingual GPT-4.5 leads
Claude Sonnet 4: 46.9 (#151), GPT-4.5: 52.7 (#78)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| LMArena Non-English | 1333 | 1413 |
| LMArena Chinese | 1351 | 1421 |
| LMArena French | 1364 | 1417 |
| LMArena German | 1331 | 1457 |
| LMArena Japanese | 1301 | 1416 |
| LMArena Korean | 1293 | 1392 |
| LMArena Russian | 1355 | 1419 |
| LMArena Spanish | 1356 | — |
Instruction Following Too close to call
Claude Sonnet 4: 71.8 (#141), GPT-4.5: 72.8 (#129)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| LMArena Instruction Following | 1375 | 1404 |
| LiveBench Instruction Following | — | 72.3% |
| IFEval | 84% | — |
Long Context GPT-4.5 leads
Claude Sonnet 4: 33.6 (#254), GPT-4.5: 40.4 (#150)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| Fiction.LiveBench | 46.9% | 63.9% |
| LMArena Longer Query | 1398 | 1406 |
Writing & Preference Too close to call
Claude Sonnet 4: 57.3 (#127), GPT-4.5: 57.1 (#129)
| Benchmark | Claude Sonnet 4 | GPT-4.5 |
|---|---|---|
| LMArena Text | 1351 | 1417 |
| LMArena Creative Writing | 1345 | 1394 |
| Short-Story Creative Writing | 81.4% | 75.6% |
| EQ-Bench Creative Writing | 1483 | 1258 |
| LMArena Multi-Turn | 1376 | 1444 |
| WildBench | 83.8% | — |
| LiveBench Language | — | 61.5% |
Frequently asked questions
Is Claude Sonnet 4 better than GPT-4.5?
Claude Sonnet 4 is the stronger model overall, scoring 40.6 to 36.9 on the Noometry Index.
Is Claude Sonnet 4 or GPT-4.5 better for coding?
Claude Sonnet 4 scores higher on coding benchmarks: 44.0 versus 41.8 in the Noometry coding category.
How many benchmarks do Claude Sonnet 4 and GPT-4.5 share?
35 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and GPT-4.5 has 42.