Model comparison
Claude Opus 5 vs GPT-4.1
Claude Opus 5 is the stronger model overall, scoring 67.8 to 35.9 on the Noometry Index. GPT-4.1 costs 2.9× less per token, which makes it the better buy when Claude Opus 5's lead doesn't matter for your workload.
Last verified . 32 shared benchmarks.
Summary
- They share 32 benchmarks with published results for both. Claude Opus 5 scores higher in 10 categories and GPT-4.1 in 0 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Claude Opus 5 leads 77.2 to 11.7.
- The biggest single-benchmark swing is ARC-AGI-1: 97.5% for Claude Opus 5 and 5.5% for GPT-4.1.
- GPT-4.1 is cheaper at $2 / $8 per million input/output tokens, against $5 / $25 for Claude Opus 5.
- GPT-4.1 accepts more context: 1.05M tokens versus 1M.
Side by side
| Claude Opus 5 | GPT-4.1 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 67.8 | 35.9 |
| Released | 2026-07-24 | 2025-04-14 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 128K | 33K |
| Input $ / M tokens | $5 | $2 |
| Output $ / M tokens | $25 | $8 |
| Results tracked | 57 | 52 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Opus 5 leads
Claude Opus 5: 67.5 (#5), GPT-4.1: 34.4 (#238)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| WeirdML | 91.8% | 39% |
| LMArena Coding | 1534 | 1391 |
| ALE-Bench | 2,165 | 558.1 |
| SWE-bench Verified | — | 48.5% |
| DeepSWE | 73.6% | — |
| FrontierCode | 53.4% | — |
| SWE-bench Verified (bash only) | — | 39.6% |
| Aider Polyglot | — | 52.4% |
| CursorBench | 46.6% | — |
| LMArena WebDev | 1691 | — |
| FrontierSWE | 52% | — |
| SciCode | 56.4% | — |
| CadEval | — | 42% |
Agentic & Tool Use Claude Opus 5 leads
Claude Opus 5: 55.6 (#1), GPT-4.1: 34.7 (#43)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| APEX-Agents | 65.8% | — |
| Berkeley Function Calling Leaderboard | — | 54% |
| OSWorld 2.0 | 31.4% | — |
| τ²-bench Banking | 48.7% | — |
| PostTrainBench | 35% | — |
| BALROG | 63.4% | — |
| GBAEval | 79.6% | — |
| GDP.pdf | 24% | — |
| Vending-Bench 2 | 11,182 | — |
Reasoning Claude Opus 5 leads
Claude Opus 5: 77.2 (#4), GPT-4.1: 11.7 (#339)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| ARC-AGI-2 | 90.4% | 0.4% |
| SimpleBench | 80.6% | 27% |
| ARC-AGI-1 | 97.5% | 5.5% |
| Chess Puzzles | 42% | 6% |
| LMArena Hard Prompts | 1526 | 1384 |
| DTBench | 97.9% | 68.3% |
| LMCA | 64.5% | 25.6% |
| Epoch Capabilities Index | 162.78 | 136.78 |
| Kagi LLM Benchmark | — | 52.3% |
| NYT Connections (extended) | 94.3% | — |
| CritPt | 29.1% | — |
| EnigmaEval | — | 2.2% |
| EBR-Bench | 45.7% | — |
| Mystery Game Puzzles | 59% | — |
| Bench to the Future 3 | 0.12 | — |
| ForecastBench | — | 61.5 |
Math Claude Opus 5 leads
Claude Opus 5: 86.2 (#8), GPT-4.1: 22.3 (#280)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 85.6% | 6% |
| OTIS Mock AIME 2024-2025 | 98.9% | 38.3% |
| LMArena Math | 1531 | 1370 |
| FrontierMath Tier 4 | 73.2% | — |
| ProofBench | 99% | — |
| Omni-MATH | — | 47.1% |
| MATH Level 5 | — | 83% |
| FrontierMath (Feb 2025 set) | — | 5.5% |
| FrontierMath Tier 4 (v1) | — | 0% |
Knowledge Claude Opus 5 leads
Claude Opus 5: 66.8 (#9), GPT-4.1: 37.1 (#160)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| GPQA Diamond | 93.9% | 66.9% |
| SimpleQA Verified | 59.9% | 31.1% |
| LMArena Expert | 1557 | 1364 |
| Humanity's Last Exam | — | 5.4% |
| MMLU-Pro | — | 81.1% |
| Vectara Hallucination Rate | — | 5.6% |
| GPQA (HELM) | — | 65.9% |
Multimodal Claude Opus 5 leads
Claude Opus 5: 50.8 (#8), GPT-4.1: 38.2 (#67)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| LMArena Vision | 1319 | 1211 |
| GeoBench | — | 72% |
| Blueprint-Bench 2 | 30.4% | — |
| Furniture Assembly | 60.8% | — |
| LMArena Document | 1516 | — |
Multilingual Claude Opus 5 leads
Claude Opus 5: 58.8 (#4), GPT-4.1: 49.4 (#133)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| LMArena Non-English | 1501 | 1370 |
| LMArena Chinese | 1574 | 1382 |
| LMArena French | 1519 | 1382 |
| LMArena German | 1524 | 1381 |
| LMArena Japanese | 1516 | 1319 |
| LMArena Korean | 1521 | 1339 |
| LMArena Russian | 1507 | 1377 |
| LMArena Spanish | 1519 | 1376 |
Instruction Following Claude Opus 5 leads
Claude Opus 5: 79.2 (#7), GPT-4.1: 71.3 (#153)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| LMArena Instruction Following | 1517 | 1367 |
| IFEval | — | 83.8% |
Long Context Claude Opus 5 leads
Claude Opus 5: 46.5 (#21), GPT-4.1: 40.0 (#163)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| LMArena Longer Query | 1515 | 1385 |
| Fiction.LiveBench | — | 63.9% |
Writing & Preference Claude Opus 5 leads
Claude Opus 5: 79.2 (#1), GPT-4.1: 57.6 (#125)
| Benchmark | Claude Opus 5 | GPT-4.1 |
|---|---|---|
| LMArena Text | 1507 | 1383 |
| LMArena Creative Writing | 1491 | 1363 |
| EQ-Bench Creative Writing | 2133 | 1420 |
| LMArena Multi-Turn | 1499 | 1398 |
| WildBench | — | 85.4% |
| EQ-Bench 4 | 1385 | — |
Frequently asked questions
Is Claude Opus 5 better than GPT-4.1?
Claude Opus 5 is the stronger model overall, scoring 67.8 to 35.9 on the Noometry Index. GPT-4.1 costs 2.9× less per token, which makes it the better buy when Claude Opus 5's lead doesn't matter for your workload.
Which is cheaper, Claude Opus 5 or GPT-4.1?
GPT-4.1 is cheaper. It lists at $2 per million input tokens and $8 per million output tokens; Claude Opus 5 lists at $5 and $25.
Is Claude Opus 5 or GPT-4.1 better for coding?
Claude Opus 5 scores higher on coding benchmarks: 67.5 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
GPT-4.1 does, with 1.05M tokens against 1M.
How many benchmarks do Claude Opus 5 and GPT-4.1 share?
32 benchmarks have published results for both models. Claude Opus 5 has 57 scored results on Noometry and GPT-4.1 has 52.