Model comparison
Claude Sonnet 4.5 vs GPT-5 Pro
GPT-5 Pro is the stronger model overall, scoring 46.2 to 44.0 on the Noometry Index. Claude Sonnet 4.5 costs 6.9× less per token, which makes it the better buy when GPT-5 Pro's lead doesn't matter for your workload.
Last verified . 12 shared benchmarks.
Summary
- They share 12 benchmarks with published results for both. Claude Sonnet 4.5 scores higher in 1 category and GPT-5 Pro in 3 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5 Pro leads 48.0 to 31.9.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 23.9% for Claude Sonnet 4.5 and 55.8% for GPT-5 Pro.
- Claude Sonnet 4.5 is cheaper at $3 / $15 per million input/output tokens, against $15 / $120 for GPT-5 Pro.
- Claude Sonnet 4.5 accepts more context: 1M tokens versus 400K.
Side by side
| Claude Sonnet 4.5 | GPT-5 Pro | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 44.0 | 46.2 |
| Released | 2025-09-29 | 2025-10-06 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 400K |
| Max output | 64K | 272K |
| Input $ / M tokens | $3 | $15 |
| Output $ / M tokens | $15 | $120 |
| Results tracked | 73 | 12 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4.5 leads
Claude Sonnet 4.5: 47.8 (#54), GPT-5 Pro: 43.4 (#77)
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| WeirdML | 47.7% | 60.4% |
| AlgoTune | 1.52 | 1.31 |
| SWE-bench Verified | 71.3% | — |
| SWE-bench Verified (bash only) | 71.4% | — |
| LMArena WebDev | 1393 | — |
| SWE-bench Multilingual | 67% | — |
| SciCode | 44.7% | — |
| GSO | 14.7% | — |
| LMArena Coding | 1489 | — |
| ALE-Bench | 796.15 | — |
Agentic & Tool Use Not comparable
Claude Sonnet 4.5: 38.1 (#30), GPT-5 Pro: —
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| Terminal-Bench | 46.5% | — |
| Berkeley Function Calling Leaderboard | 73.2% | — |
| GDPval | 42.5% | — |
| Remote Labor Index | 2.1% | — |
| τ²-bench Airline | 72% | — |
| τ²-bench Banking | 25.3% | — |
| τ²-bench Retail | 72.4% | — |
| τ²-bench Telecom | 84.9% | — |
| Cybench | 60% | — |
| DeepResearch Bench | 52.6% | — |
| OSWorld | 62.9% | — |
| LMArena Search | 1159 | — |
| METR Time Horizons | 67.4% | — |
| Vending-Bench 2 | 3,839 | — |
Reasoning GPT-5 Pro leads
Claude Sonnet 4.5: 26.9 (#113), GPT-5 Pro: 39.2 (#52)
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| ARC-AGI-2 | 13.6% | 18.3% |
| SimpleBench | 54.3% | 61.6% |
| Kagi LLM Benchmark | 57.9% | 76.8% |
| ARC-AGI-1 | 63.7% | 70.2% |
| EnigmaEval | 6% | 18.8% |
| Epoch Capabilities Index | 146.84 | 150.26 |
| NYT Connections (extended) | 37.3% | — |
| CritPt | 1.1% | — |
| Chess Puzzles | 12% | — |
| EBR-Bench | 2.4% | — |
| LMArena Hard Prompts | 1462 | — |
| Mystery Game Puzzles | 17% | — |
| DTBench | 83.2% | — |
| LMCA | 38.8% | — |
| ForecastBench | 61.9 | — |
Math GPT-5 Pro leads
Claude Sonnet 4.5: 31.9 (#190), GPT-5 Pro: 48.0 (#53)
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| FrontierMath (Tiers 1-3) | 23.9% | 55.8% |
| FrontierMath Tier 4 | 2.4% | 19.5% |
| FrontierMath Tier 4 (v1) | 4.2% | 14.6% |
| OTIS Mock AIME 2024-2025 | 77.8% | — |
| ProofBench | 19% | — |
| Omni-MATH | 55.3% | — |
| LMArena Math | 1448 | — |
| MATH Level 5 | 97.7% | — |
| FrontierMath (Feb 2025 set) | 15.2% | — |
Knowledge GPT-5 Pro leads
Claude Sonnet 4.5: 48.4 (#70), GPT-5 Pro: 56.2 (#39)
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| Humanity's Last Exam | 13.7% | 31.6% |
| GPQA Diamond | 82.3% | — |
| SimpleQA Verified | 30.7% | — |
| MMLU-Pro | 86.9% | — |
| Vectara Hallucination Rate | 12% | — |
| GPQA (HELM) | 68.6% | — |
| LMArena Expert | 1483 | — |
Multimodal Not comparable
Claude Sonnet 4.5: 34.1 (#84), GPT-5 Pro: —
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| VPCT | 39.8% | — |
| LMArena Document | 1450 | — |
Multilingual Not comparable
Claude Sonnet 4.5: 53.6 (#65), GPT-5 Pro: —
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| LMArena Non-English | 1425 | — |
| LMArena Chinese | 1459 | — |
| LMArena French | 1459 | — |
| LMArena German | 1427 | — |
| LMArena Japanese | 1390 | — |
| LMArena Korean | 1403 | — |
| LMArena Russian | 1436 | — |
| LMArena Spanish | 1457 | — |
Instruction Following Not comparable
Claude Sonnet 4.5: 75.2 (#74), GPT-5 Pro: —
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| IFEval | 85% | — |
| LMArena Instruction Following | 1459 | — |
Long Context Not comparable
Claude Sonnet 4.5: 45.2 (#45), GPT-5 Pro: —
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| LMArena Longer Query | 1476 | — |
Writing & Preference Not comparable
Claude Sonnet 4.5: 66.7 (#33), GPT-5 Pro: —
| Benchmark | Claude Sonnet 4.5 | GPT-5 Pro |
|---|---|---|
| LMArena Text | 1439 | — |
| LMArena Creative Writing | 1442 | — |
| EQ-Bench Creative Writing | 1678 | — |
| WildBench | 85.4% | — |
| LMArena Multi-Turn | 1465 | — |
Frequently asked questions
Is Claude Sonnet 4.5 better than GPT-5 Pro?
GPT-5 Pro is the stronger model overall, scoring 46.2 to 44.0 on the Noometry Index. Claude Sonnet 4.5 costs 6.9× less per token, which makes it the better buy when GPT-5 Pro's lead doesn't matter for your workload.
Which is cheaper, Claude Sonnet 4.5 or GPT-5 Pro?
Claude Sonnet 4.5 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; GPT-5 Pro lists at $15 and $120.
Is Claude Sonnet 4.5 or GPT-5 Pro better for coding?
Claude Sonnet 4.5 scores higher on coding benchmarks: 47.8 versus 43.4 in the Noometry coding category.
Which has the bigger context window?
Claude Sonnet 4.5 does, with 1M tokens against 400K.
How many benchmarks do Claude Sonnet 4.5 and GPT-5 Pro share?
12 benchmarks have published results for both models. Claude Sonnet 4.5 has 73 scored results on Noometry and GPT-5 Pro has 12.