Model comparison
Claude 3.7 Sonnet vs GPT-5
GPT-5 is the stronger model overall, scoring 50.9 to 39.4 on the Noometry Index.
Last verified . 46 shared benchmarks.
Summary
- They share 46 benchmarks with published results for both. Claude 3.7 Sonnet scores higher in 1 category and GPT-5 in 9 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5 leads 38.3 to 18.7.
- The biggest single-benchmark swing is ARC-AGI-1: 28.6% for Claude 3.7 Sonnet and 65.7% for GPT-5.
Side by side
| Claude 3.7 Sonnet | GPT-5 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 39.4 | 50.9 |
| Released | 2025-02-24 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | — | 400K |
| Max output | — | 128K |
| Input $ / M tokens | — | $1.25 |
| Output $ / M tokens | — | $10 |
| Results tracked | 58 | 69 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 leads
Claude 3.7 Sonnet: 41.1 (#120), GPT-5: 50.7 (#39)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| SWE-bench Verified | 61% | 73.6% |
| SWE-bench Verified (bash only) | 52.8% | 65% |
| Aider Polyglot | 64.9% | 88% |
| GSO | 3.8% | 6.9% |
| LMArena Coding | 1361 | 1436 |
| LMArena WebDev | — | 1417 |
| SciCode | — | 42.9% |
| WeirdML | — | 60.7% |
| LiveBench Coding | 74.5% | — |
| CadEval | 54% | — |
| ALE-Bench | — | 1,162 |
| AlgoTune | — | 1.67 |
Agentic & Tool Use Too close to call
Claude 3.7 Sonnet: 33.9 (#46), GPT-5: 33.0 (#53)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| DeepResearch Bench | 43.6% | 49.6% |
| METR Time Horizons | 60% | 69.6% |
| Terminal-Bench | — | 49.6% |
| GDPval | — | 34.8% |
| Remote Labor Index | — | 1.7% |
| TheAgentCompany | 30.9% | — |
| Cybench | 20% | — |
| OSWorld | 35.8% | — |
| BALROG | — | 32.8% |
| LMArena Search | — | 1133 |
Reasoning GPT-5 leads
Claude 3.7 Sonnet: 18.7 (#252), GPT-5: 38.3 (#55)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| ARC-AGI-2 | 0.9% | 9.9% |
| SimpleBench | 46.4% | 56.7% |
| ARC-AGI-1 | 28.6% | 65.7% |
| EnigmaEval | 4.2% | 10.5% |
| LMArena Hard Prompts | 1333 | 1415 |
| Epoch Capabilities Index | 141.16 | 150 |
| ForecastBench | 61.8 | 61.4 |
| Kagi LLM Benchmark | — | 72.7% |
| CritPt | — | 12.6% |
| Chess Puzzles | — | 37% |
| EBR-Bench | — | 12.7% |
| LiveBench Reasoning | 87.8% | — |
| Mystery Game Puzzles | — | 23% |
| DTBench | — | 90.7% |
| LiveBench Data Analysis | 74% | — |
| LMCA | — | 40% |
| LiveBench | 76.1% | — |
Math GPT-5 leads
Claude 3.7 Sonnet: 35.7 (#137), GPT-5: 54.6 (#36)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 57.8% | 91.4% |
| Omni-MATH | 33% | 64.7% |
| LMArena Math | 1336 | 1406 |
| MATH Level 5 | 91.2% | 98.1% |
| FrontierMath (Feb 2025 set) | 4.1% | 32.4% |
| FrontierMath (Tiers 1-3) | — | 55.4% |
| FrontierMath Tier 4 | — | 22% |
| ProofBench | — | 18% |
| LiveBench Math | 79% | — |
| FrontierMath Tier 4 (v1) | — | 12.5% |
Knowledge GPT-5 leads
Claude 3.7 Sonnet: 39.7 (#117), GPT-5: 56.5 (#37)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| GPQA Diamond | 79.7% | 86.2% |
| Humanity's Last Exam | 8% | 25.3% |
| MMLU-Pro | 78.4% | 86.3% |
| Confabulations | 14.7% | 10.3% |
| GPQA (HELM) | 60.8% | 79.2% |
| LMArena Expert | 1321 | 1420 |
| SimpleQA Verified | — | 50.1% |
| Vectara Hallucination Rate | — | 14.7% |
Multimodal GPT-5 leads
Claude 3.7 Sonnet: 33.1 (#90), GPT-5: 46.1 (#12)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| LMArena Vision | 1169 | 1232 |
| GeoBench | 68% | 81% |
| VPCT | 39% | 66% |
| SpatialViz-Bench | 33.9% | — |
Multilingual GPT-5 leads
Claude 3.7 Sonnet: 44.3 (#174), GPT-5: 51.5 (#105)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| LMArena Non-English | 1296 | 1397 |
| LMArena Chinese | 1299 | 1423 |
| LMArena French | 1302 | 1409 |
| LMArena German | 1301 | 1415 |
| LMArena Japanese | 1268 | 1411 |
| LMArena Korean | 1249 | 1361 |
| LMArena Russian | 1311 | 1406 |
| LMArena Spanish | 1299 | 1398 |
Instruction Following Too close to call
Claude 3.7 Sonnet: 73.1 (#120), GPT-5: 74.0 (#107)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| IFEval | 83.4% | 87.5% |
| LMArena Instruction Following | 1352 | 1388 |
| LiveBench Instruction Following | 81.3% | — |
Long Context GPT-5 leads
Claude 3.7 Sonnet: 50.2 (#10), GPT-5: 69.5 (#2)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| Fiction.LiveBench | 83.3% | 97.2% |
| LMArena Longer Query | 1373 | 1399 |
Writing & Preference GPT-5 leads
Claude 3.7 Sonnet: 54.6 (#145), GPT-5: 63.5 (#63)
| Benchmark | Claude 3.7 Sonnet | GPT-5 |
|---|---|---|
| LMArena Text | 1314 | 1406 |
| LMArena Creative Writing | 1332 | 1365 |
| Short-Story Creative Writing | 81.1% | 86% |
| EQ-Bench Creative Writing | 1412 | 1627 |
| WildBench | 81.4% | 85.7% |
| LMArena Multi-Turn | 1339 | 1426 |
| LiveBench Language | 59.9% | — |
Frequently asked questions
Is Claude 3.7 Sonnet better than GPT-5?
GPT-5 is the stronger model overall, scoring 50.9 to 39.4 on the Noometry Index.
Is Claude 3.7 Sonnet or GPT-5 better for coding?
GPT-5 scores higher on coding benchmarks: 50.7 versus 41.1 in the Noometry coding category.
How many benchmarks do Claude 3.7 Sonnet and GPT-5 share?
46 benchmarks have published results for both models. Claude 3.7 Sonnet has 58 scored results on Noometry and GPT-5 has 69.