Model comparison
Claude Sonnet 4.5 vs GPT-5
GPT-5 is the stronger model overall, scoring 50.9 to 44.1 on the Noometry Index.
Last verified . 62 shared benchmarks.
Summary
- They share 62 benchmarks with published results for both. Claude Sonnet 4.5 scores higher in 4 categories and GPT-5 in 6 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in long context, where GPT-5 leads 69.5 to 45.2.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 23.9% for Claude Sonnet 4.5 and 55.4% for GPT-5.
- GPT-5 is cheaper at $1.25 / $10 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.5.
- GPT-5 accepts more context: 400K tokens versus 200K.
Side by side
| Claude Sonnet 4.5 | GPT-5 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 44.1 | 50.9 |
| Released | 2025-09-29 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 400K |
| Max output | 64K | 128K |
| Input $ / M tokens | $3 | $1.25 |
| Output $ / M tokens | $15 | $10 |
| Results tracked | 73 | 69 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 leads
Claude Sonnet 4.5: 47.3 (#61), GPT-5: 50.3 (#47)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| SWE-bench Verified | 71.3% | 73.6% |
| SWE-bench Verified (bash only) | 71.4% | 65% |
| LMArena WebDev | 1393 | 1418 |
| SciCode | 44.7% | 42.9% |
| GSO | 14.7% | 6.9% |
| WeirdML | 47.7% | 60.7% |
| LMArena Coding | 1489 | 1436 |
| ALE-Bench | 796.15 | 1,162 |
| AlgoTune | 1.52 | 1.67 |
| Aider Polyglot | — | 88% |
| SWE-bench Multilingual | 67% | — |
Agentic & Tool Use Claude Sonnet 4.5 leads
Claude Sonnet 4.5: 38.3 (#32), GPT-5: 33.1 (#56)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| Terminal-Bench | 46.5% | 49.6% |
| GDPval | 42.5% | 34.8% |
| Remote Labor Index | 2.1% | 1.7% |
| DeepResearch Bench | 52.6% | 49.6% |
| LMArena Search | 1159 | 1133 |
| METR Time Horizons | 67.4% | 69.6% |
| Berkeley Function Calling Leaderboard | 73.2% | — |
| τ²-bench Airline | 72% | — |
| τ²-bench Banking | 25.3% | — |
| τ²-bench Retail | 72.4% | — |
| τ²-bench Telecom | 84.9% | — |
| Cybench | 60% | — |
| OSWorld | 62.9% | — |
| BALROG | — | 32.8% |
| Vending-Bench 2 | 3,839 | — |
Reasoning GPT-5 leads
Claude Sonnet 4.5: 26.9 (#125), GPT-5: 38.3 (#64)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| ARC-AGI-2 | 13.6% | 9.9% |
| SimpleBench | 54.3% | 56.7% |
| Kagi LLM Benchmark | 57.9% | 72.7% |
| ARC-AGI-1 | 63.7% | 65.7% |
| CritPt | 1.1% | 12.6% |
| Chess Puzzles | 12% | 37% |
| EnigmaEval | 6% | 10.5% |
| EBR-Bench | 2.4% | 12.7% |
| LMArena Hard Prompts | 1462 | 1416 |
| Mystery Game Puzzles | 17% | 23% |
| DTBench | 83.2% | 90.7% |
| LMCA | 38.8% | 40% |
| Epoch Capabilities Index | 146.84 | 150 |
| ForecastBench | 61.9 | 61.4 |
| NYT Connections (extended) | 37.3% | — |
Math GPT-5 leads
Claude Sonnet 4.5: 32.3 (#216), GPT-5: 55.0 (#44)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 23.9% | 55.4% |
| FrontierMath Tier 4 | 2.4% | 22% |
| OTIS Mock AIME 2024-2025 | 77.8% | 91.4% |
| ProofBench | 19% | 18% |
| Omni-MATH | 55.3% | 64.7% |
| LMArena Math | 1449 | 1407 |
| MATH Level 5 | 97.7% | 98.1% |
| FrontierMath (Feb 2025 set) | 15.2% | 32.4% |
| FrontierMath Tier 4 (v1) | 4.2% | 12.5% |
Knowledge GPT-5 leads
Claude Sonnet 4.5: 48.4 (#76), GPT-5: 56.6 (#43)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| GPQA Diamond | 82.3% | 86.2% |
| Humanity's Last Exam | 13.7% | 25.3% |
| SimpleQA Verified | 30.7% | 50.1% |
| MMLU-Pro | 86.9% | 86.3% |
| Vectara Hallucination Rate | 12% | 14.7% |
| GPQA (HELM) | 68.6% | 79.2% |
| LMArena Expert | 1482 | 1419 |
| Confabulations | — | 10.3% |
Multimodal GPT-5 leads
Claude Sonnet 4.5: 34.8 (#89), GPT-5: 46.8 (#13)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| VPCT | 39.8% | 66% |
| LMArena Vision | — | 1232 |
| GeoBench | — | 81% |
| LMArena Document | 1450 | — |
Multilingual Claude Sonnet 4.5 leads
Claude Sonnet 4.5: 53.4 (#69), GPT-5: 51.4 (#110)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| LMArena Non-English | 1425 | 1397 |
| LMArena Chinese | 1459 | 1422 |
| LMArena French | 1458 | 1410 |
| LMArena German | 1427 | 1416 |
| LMArena Japanese | 1390 | 1409 |
| LMArena Korean | 1403 | 1360 |
| LMArena Russian | 1437 | 1406 |
| LMArena Spanish | 1457 | 1399 |
Instruction Following Claude Sonnet 4.5 leads
Claude Sonnet 4.5: 75.0 (#78), GPT-5: 73.8 (#113)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| IFEval | 85% | 87.5% |
| LMArena Instruction Following | 1459 | 1388 |
Long Context GPT-5 leads
Claude Sonnet 4.5: 45.2 (#46), GPT-5: 69.5 (#2)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| LMArena Longer Query | 1476 | 1399 |
| Fiction.LiveBench | — | 97.2% |
Writing & Preference Claude Sonnet 4.5 leads
Claude Sonnet 4.5: 66.5 (#34), GPT-5: 63.4 (#65)
| Benchmark | Claude Sonnet 4.5 | GPT-5 |
|---|---|---|
| LMArena Text | 1439 | 1406 |
| LMArena Creative Writing | 1442 | 1365 |
| EQ-Bench Creative Writing | 1678 | 1627 |
| WildBench | 85.4% | 85.7% |
| LMArena Multi-Turn | 1465 | 1426 |
| Short-Story Creative Writing | — | 86% |
Frequently asked questions
Is Claude Sonnet 4.5 better than GPT-5?
GPT-5 is the stronger model overall, scoring 50.9 to 44.1 on the Noometry Index.
Which is cheaper, Claude Sonnet 4.5 or GPT-5?
GPT-5 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; Claude Sonnet 4.5 lists at $3 and $15.
Is Claude Sonnet 4.5 or GPT-5 better for coding?
GPT-5 scores higher on coding benchmarks: 50.3 versus 47.3 in the Noometry coding category.
Which has the bigger context window?
GPT-5 does, with 400K tokens against 200K.
How many benchmarks do Claude Sonnet 4.5 and GPT-5 share?
62 benchmarks have published results for both models. Claude Sonnet 4.5 has 73 scored results on Noometry and GPT-5 has 69.